Renewable EnergyEdited by T J HammonsEdited by T J HammonsRenewable Energy is energy generated from natural resources such as sunlight, wind, rain, tides and geothermal heat which are naturally replenished. In 2008, about 18% of global final energy consumption came from renewables, with 13% coming from traditional biomass, such as wood burning. Hydroelectricity was the next largest renewable source, providing 3% (15% of global electricity generation), followed by solar hot water/heating, which contributed with 1.3%. Modern technologies, such as geothermal energy, wind power, solar power, and ocean energy together provided some 0.8% of final energy consumption. The book provides a forum for dissemination and exchange of up to date scientific information on theoretical, generic and applied areas of knowledge. The topics deal with new devices and circuits for energy systems, photovoltaic and solar thermal, wind energy systems, tidal and wave energy, fuel cell systems, bio energy and geo-energy, sustainable energy resources and systems, energy storage systems, energy market management and economics, off-grid isolated energy systems, energy in transportation systems, energy resources for portable electronics, intelligent energy power transmission, distribution and interconnectors, energy efficient utilization, environmental issues, energy harvesting, nanotechnology in energy, policy issues on renewable energy, building design, power electronics in energy conversion, new materials for energy resources, and RF and magnetic field energy devices.Photo by DiyanaDimitrova / iStockISBN 978-953-7619-52-7Renewable EnergyRenewable EnergyEdited byT J HammonsIn-Tech intechweb.orgRenewable EnergyEdited byT J HammonsIn-Tech intechweb.orgRenewable Energyhttp://dx.doi.org/10.5772/45752Edited by T J HammonsContributorsReggie Davidrajuh, Joao Figueiredo, Lyle Jenkins, Samuel Jupe, Andrea Michiorri, Philip Taylor, Sergio Faias, Jorge Sousa, Rui Castro, Emilio Lorenzani, Giovanni Franceschini, Alberto Bellini Bellini, Carla Tassoni, Athula Rajapakse, Dharshana Muthumuni, Nuwan Perera, Marco Mauri, Ali Eltamaly, Ilse Cervantes, Angelica Mendoza-Torres, Francisco Perez-Pinal, Branislav Dobrucky, Michal Pokorný, Mariana Benova, Yun Seng Lim, Siong Lee Koh, Levente Tamas, Zoltan Szekely, Mamadou Lamine Doumbia, Kodjo Agbossou, Firuz Zare, Dorin Bica, Cristian Dragoş Dumitru, Adrian Gligor, Adrian-Vasile Duka, Bjorn Bakken, Florin Iov, Frede Blaabjerg, Nazih Moubayed, Ali El-Ali, Rachid Outbib, Joao Martins, Carmen M. Rangel, Antonio Joyce, Joao Sotomayor, Armando Pires, Jose Pelegri-Sebastia, Jan Machacek, Zdenek Prochazka, Jiri Drapela, Eleonora Darie, Costin Cepisca, Emanuel Darie, Thomas Hammons, Roque Mesquita Brandao, Beleza Carvalho, F. Maciel-Barbosa, Francois Vallee, Jon Are Suul© The Editor(s) and the Author(s) 2009The moral rights of the and the author(s) have been asserted.All rights to the book as a whole are reserved by INTECH. The book as a whole (compilation) cannot be reproduced, distributed or used for commercial or non-commercial purposes without INTECH’s written permission. Enquiries concerning the use of the book should be directed to INTECH rights and permissions department (permissions@intechopen.com).Violations are liable to prosecution under the governing Copyright Law.Individual chapters of this publication are distributed under the terms of the Creative Commons Attribution 3.0 Unported License which permits commercial use, distribution and reproduction of the individual chapters, provided the original author(s) and source publication are appropriately acknowledged. If so indicated, certain images may not be included under the Creative Commons license. In such cases users will need to obtain permission from the license holder to reproduce the material. More details and guidelines concerning content reuse and adaptation can be foundat http://www.intechopen.com/copyright-policy.html.NoticeStatements and opinions expressed in the chapters are these of the individual contributors and not necessarily those of the editors or publisher. No responsibility is accepted for the accuracy of information contained in the published chapters. The publisher assumes no responsibility for any damage or injury to persons or property arising out of the use of any materials, instructions, methods or ideas contained in the book.First published in Croatia, 2009 by INTECH d.o.o.eBook (PDF) Published by IN TECH d.o.o.Place and year of publication of eBook (PDF): Rijeka, 2019.IntechOpen is the global imprint of IN TECH d.o.o.Printed in CroatiaLegal deposit, Croatia: National and University Library in ZagrebAdditional hard and PDF copies can be obtained from orders@intechopen.comRenewable EnergyEdited by T J Hammonsp. cm.ISBN 978-953-7619-52-7eBook (PDF) ISBN 978-953-51-6416-6Selection of our books indexed in the Book Citation Index in Web of Science™ Core Collection (BKCI)Interested in publishing with us? Contact book.department@intechopen.comNumbers displayed above are based on latest data collected. For more information visit www.intechopen.com4,000+ Open access books available151Countries delivered to12.2%Contributors from top 500 universitiesOur authors are among theTop 1%most cited scientists116,000+International authors and editors120M+ DownloadsWe are IntechOpen,the world’s leading publisher of Open Access booksBuilt by scientists, for scientists Meet the editorThomas James Hammons (Fellow IEEE 1996) received the B.Sc. degree in Engineering (1st Class Honors), and the DIC, and Ph.D. degrees from Imperial College, Lon-don, UK He was Professor of Electrical and Computer Engineering at McMaster University, Hamilton, On-tario, Canada in 1978-1979. He is the author/co-author of over 440 scientific articles and papers on electrical power engineering and is Editor of a book on Renewable Energy that was published by INTECH in December 2009. He has lectured extensively in North America, Africa, Asia, and both in Eastern and Western Europe. Dr Hammons is Past Chair of the United Kingdom and Republic of Ireland (UKRI) Section IEEE and Past Chair of International Practices for Energy Development and Power Generation of IEEE. He is also a Past Chair of the IEEE PES Task Force on harmonizing power-engineering standards worldwide and Past Permanent Secretary of the International Universities Power Engineering Conference. He is a Chartered Engineer (CEng) and a registered European Engineer in the Federation of National Engineering Associations in Europe.VPrefaceOur goal in preparing this book was to discuss and publish new discoveries and improvements, innovative ideas and concepts, as well as novel and further applications and business models which are related to the field of Renewable Energy. Renewable Energy is energy generated from natural resources—such as sunlight, wind, rain, tides and geothermal heat—which are naturally replenished. In 2008, about 18% of global final energy consumption came from renewables, with 13% coming from traditional biomass, such as wood burning. Hydroelectricity was the next largest renewable source, providing 3% (15% of global electricity generation), followed by solar hot water/heating, which contributed 1.3%. Modern technologies, such as geothermal energy, wind power, solar power, and ocean energy together provided some 0.8% of final energy consumption.Alternative energy includes all sources and technologies that minimize environmental impacts relative to conventional hydrocarbon resources and economic issues related to fossil fuel resources. Fuel cells and natural gas might be alternatives to coal or nuclear power. Throughout the book, the fundamentals of the technologies related to integration of such alternative and renewable energy sources are reviewed and described with authority, skill, and from critical engineering aspects for the end user of energy. Climate change concerns coupled with oil prices with its uncertainty and increasing government support is driving increasing renewable energy legislation, incentives and commercialization. Investment capital flowing into renewable energy climbed from $80 billion (US) in 2005 to $100 billion in 2006. The book provides the forum for dissemination and exchange of up-to-date scientific information on theoretical, generic and applied areas of knowledge.The topics deal with new devices and circuits for energy systems, photovoltaic and solar thermal, wind energy systems, tidal and wave energy, fuel cell systems, bio energy and geo-energy, sustainable energy resources and systems, energy storage systems, energy market management and economics, off-grid isolated energy systems, energy in transportation systems, energy resources for portable electronics, intelligent energy power transmission, distribution and inter-connectors, energy efficient utilization, environmental issues, energy harvesting, nanotechnology in energy, policy issues on renewable energy, building design, power electronics in energy conversion, new materials for energy resources, and RF and magnetic field energy devices. Open Access is a new direction in academic publishing where all chapters are available for full free access online. The book is published as open access fully searchable by anyone anywhere. VPrefaceOur goal in preparing this book was to discuss and publish new discoveries and improvements, innovative ideas and concepts, as well as novel and further applications and business models which are related to the field of Renewable Energy. Renewable Energy is energy generated from natural resources—such as sunlight, wind, rain, tides and geothermal heat—which are naturally replenished. In 2008, about 18% of global final energy consumption came from renewables, with 13% coming from traditional biomass, such as wood burning. Hydroelectricity was the next largest renewable source, providing 3% (15% of global electricity generation), followed by solar hot water/heating, which contributed 1.3%. Modern technologies, such as geothermal energy, wind power, solar power, and ocean energy together provided some 0.8% of final energy consumption.Alternative energy includes all sources and technologies that minimize environmental impacts relative to conventional hydrocarbon resources and economic issues related to fossil fuel resources. Fuel cells and natural gas might be alternatives to coal or nuclear power. Throughout the book, the fundamentals of the technologies related to integration of such alternative and renewable energy sources are reviewed and described with authority, skill, and from critical engineering aspects for the end user of energy. Climate change concerns coupled with oil prices with its uncertainty and increasing government support is driving increasing renewable energy legislation, incentives and commercialization. Investment capital flowing into renewable energy climbed from $80 billion (US) in 2005 to $100 billion in 2006. The book provides the forum for dissemination and exchange of up-to-date scientific information on theoretical, generic and applied areas of knowledge.The topics deal with new devices and circuits for energy systems, photovoltaic and solar thermal, wind energy systems, tidal and wave energy, fuel cell systems, bio energy and geo-energy, sustainable energy resources and systems, energy storage systems, energy market management and economics, off-grid isolated energy systems, energy in transportation systems, energy resources for portable electronics, intelligent energy power transmission, distribution and inter-connectors, energy efficient utilization, environmental issues, energy harvesting, nanotechnology in energy, policy issues on renewable energy, building design, power electronics in energy conversion, new materials for energy resources, and RF and magnetic field energy devices. Open Access is a new direction in academic publishing where all chapters are available for full free access online. The book is published as open access fully searchable by anyone anywhere. XWe believe that immediate, worldwide, barrier-free, open access to the full text of research articles is in the best interests of the scientific community. Free online availability substantially increases an article’s impact. The mean number of citations to offline articles has been shown to be 2.5~3.0 % smaller. The book is also available in printed edition (hardcopy} and has been distributed to major university and learned societies libraries, etc. free worldwide.The timely initiatives taken by the authors in this book to cover closely renewable energy is applauded. I am confident that this work will contribute to our better understanding of how to integrate renewable energy sources into our electricity grids for commercial, domestic and industrial applications. I strongly recommend this work to a wide audience, including environments, engineering educators, students, industrialists, consultants, and those concerned in reducing greenhouse emissions that is affecting our planet.December 2009T J HammonsUniversity of Glasgow, UKVIIContentsPreface V1. A Model for Greener Power Generation for North-east Sri Lanka based onStand-alone Renewable Energy Systems 001Reggie Davidrajuh2. Automatic Sun-Tracker System for Photo-Voltaic Plants 017Joao M. G. Figueiredo3. Development of Space-Based Solar Power 027Lyle M. Jenkins4. Increasing the energy yield of generation from new and renewable energyresources 037Samuel C. E. Jupe, Andrea Michiorri and Philip C. Taylor5. Embedded Energy Storage Systems in the Power Grid for Renewable EnergySources Integration 063Sérgio Faias, Jorge Sousa and Rui Castro6. Single-Phase Grid Connected Converters for Photovoltaic Plants 089Emilio Lorenzani, Giovanni Franceschini, Alberto Bellini and Carla Tassoni7. Grid Integration of Renewable Energy Systems 109Athula Rajapakse, Dharshana Muthumuni and Nuwan Perera8. Hardware in the loop simulation of renewable distributed generation systems 133Marco Mauri9. Harmonics Reduction Techniques in Renewable Energy Interfacing Converters 153Ali M. Eltamaly, Ph.D10. Hybrid Control of DC-DC Power Converters 173Ilse Cervantes, Francisco J. Perez-Pinal and Angelica Mendoza-Torres11. Interaction of Renewable Energy Source and Power Supply Network 197Branislav Dobrucký, Michal Pokorný and Mariana Beňová12. Marine Tidal Current Electric Power Generation: State of Art and Current Status 211Yun Seng. Lim and Siong Lee. KohVIWe believe that immediate, worldwide, barrier-free, open access to the full text of research articles is in the best interests of the scientific community. Free online availability substantially increases an article’s impact. The mean number of citations to offline articles has been shown to be 2.5~3.0 % smaller. The book is also available in printed edition (hardcopy} and has been distributed to major university and learned societies libraries, etc. free worldwide.The timely initiatives taken by the authors in this book to cover closely renewable energy is applauded. I am confident that this work will contribute to our better understanding of how to integrate renewable energy sources into our electricity grids for commercial, domestic and industrial applications. I strongly recommend this work to a wide audience, including environments, engineering educators, students, industrialists, consultants, and those concerned in reducing greenhouse emissions that is affecting our planet.December 2009T J HammonsUniversity of Glasgow, UKVIIContentsPreface IX1. A Model for Greener Power Generation for North-east Sri Lanka based onStand-alone Renewable Energy Systems 001Reggie Davidrajuh2. Automatic Sun-Tracker System for Photo-Voltaic Plants 017Joao M. G. Figueiredo3. Development of Space-Based Solar Power 027Lyle M. Jenkins4. Increasing the energy yield of generation from new and renewable energyresources 037Samuel C. E. Jupe, Andrea Michiorri and Philip C. Taylor5. Embedded Energy Storage Systems in the Power Grid for Renewable EnergySources Integration 063Sérgio Faias, Jorge Sousa and Rui Castro6. Single-Phase Grid Connected Converters for Photovoltaic Plants 089Emilio Lorenzani, Giovanni Franceschini, Alberto Bellini and Carla Tassoni7. Grid Integration of Renewable Energy Systems 109Athula Rajapakse, Dharshana Muthumuni and Nuwan Perera8. Hardware in the loop simulation of renewable distributed generation systems 133Marco Mauri9. Harmonics Reduction Techniques in Renewable Energy Interfacing Converters 153Ali M. Eltamaly, Ph.D10. Hybrid Control of DC-DC Power Converters 173Ilse Cervantes, Francisco J. Perez-Pinal and Angelica Mendoza-Torres11. Interaction of Renewable Energy Source and Power Supply Network 197Branislav Dobrucký, Michal Pokorný and Mariana Beňová12. Marine Tidal Current Electric Power Generation: State of Art and Current Status 211Yun Seng. Lim and Siong Lee. KohXII13. Modelling and Simulation of an Induction Drive with Application to a Small WindTurbine Generator 227Levente TAMAS and Zoltan SZEKELY14. Photovoltaic/Wind Energy System with Hydrogen Storage 249Mamadou Lamine Doumbia and Kodjo Agbossou15. Multilevel Converters in Renewable Energy Systems 271Alireza Nami and Firuz Zare16. Isolated hybrid solar-wind-hydro renewable energy systems 297Dorin Bică, Cristian Dragoş Dumitru, Adrian Gligor, Adrian-Vasile Duka17. Planning of Distributed Energy Systems with Parallel Infrastructures: A Case study 317Bjorn H. Bakken18. Power Electronics Control of Wind Energy in Distributed Power Systems 333Florin Iov and Frede Blaabjerg19. Renewable Energy in Lebanon 365Nazih Moubayed, Ali El-Ali and Rachid Outbib20. RenH2 – A Stand-Alone Sustainable Renewable Energy System 375João Martins, Carmen M. Rangel, António Joyce, João Sotomayor, Armando Pires, Rui Castro21. Solar Power Source for autonomous sensors 401José Pelegrí-Sebastiá, Rafael Lajara Vizcaíno & Jorge Alberola Lluch22. The Temperature Dependant Efficiency of Photovoltaic Modules - a long termevaluation of experimental measurements 415Jan Machacek, Zdenek Prochazka and Jiri Drapela23. The use of Switched Reluctance Generator in Wind Energy Applications 447Eleonora Darie, Costin Cepişcă and Emanuel Darie24. Tidal Energy Technologies: Currents, Wave and Offshore Wind Power in theUnited Kingdom, Europe and North America 463T. J. Hammons25. Wind Energy Technology 505R. Mesquita Brandão, J. Beleza Carvalho & F. P. Maciel Barbosa26. Wind Generation Modelling for the Management of Electrical TransmissionSystems 531François Vallée27. Variable speed pumped storage hydropower plants for integration of wind powerin isolated power systems 553Jon Are SuulA Model for Greener Power Generation for North-eastSri Lanka based on Stand-alone Renewable Energy Systems 1A Model for Greener Power Generation for North-east Sri Lanka basedon Stand-alone Renewable Energy SystemsReggie DavidrajuhX A Model for Greener Power Generation for North-east Sri Lanka based on Stand-alone Renewable Energy Systems Reggie Davidrajuh University of StavangerNorway1. IntroductionIn Northern and Eastern Sri Lanka (NE-SL), electricity is not available in most of the places due to the civil war. There is practically no electricity production and distribution system available or planned. Since Sri Lanka, as a whole, is suffering from power-shortage, it will be not possible for her to supply electricity to NE-SL, by the time the civil war comes to an end. An alternative generation and distribution system should, therefore, be designed to supply electricity to NE-SL.The need for electric power may be analyzed by a fairly intensive and deep study, which invariably requires life cycle analysis of the society in terms of energy use and conservation,industrial and household development, population and industry distribution, export andimport of electricity, etc. In contemporary NE-SL, however, the very basics in a societal infrastructure such as industry, highways and education system, are either non-existent or in primitive form. And due to the on-going war in this region, important data for electricitypower sector decision-making (such as income and purchasing power of the population, planned transportation and highway infrastructure, environmentally sensitive areas, meteorological data, etc.) are either not available or inadequate. The objective of this chapter is to demonstrate that to achieve a sustainable electricitygeneration for NE-SL, in addition to utilizing the abundant intermittent resources NE-SL has, namely solar and wind, biomass production must be given utmost priority. The scope of this chapter is limited to design of a mixture of power plants in the cogeneration system, not involving the other aspects of electricity sector such as transmission and distribution.Limitation of this work: due to the ongoing civil war in the region, there is no recent dataavailable about energy production and consumption in this region; the data used in thischapter is from 1994 to 1996. 1VIII13. Modelling and Simulation of an Induction Drive with Application to a Small WindTurbine Generator 227Levente TAMAS and Zoltan SZEKELY14. Photovoltaic/Wind Energy System with Hydrogen Storage 249Mamadou Lamine Doumbia and Kodjo Agbossou15. Multilevel Converters in Renewable Energy Systems 271Alireza Nami and Firuz Zare16. Isolated hybrid solar-wind-hydro renewable energy systems 297Dorin Bică, Cristian Dragoş Dumitru, Adrian Gligor, Adrian-Vasile Duka17. Planning of Distributed Energy Systems with Parallel Infrastructures: A Case study 317Bjorn H. Bakken18. Power Electronics Control of Wind Energy in Distributed Power Systems 333Florin Iov and Frede Blaabjerg19. Renewable Energy in Lebanon 365Nazih Moubayed, Ali El-Ali and Rachid Outbib20. RenH2 – A Stand-Alone Sustainable Renewable Energy System 375João Martins, Carmen M. Rangel, António Joyce, João Sotomayor, Armando Pires, Rui Castro21. Solar Power Source for autonomous sensors 401José Pelegrí-Sebastiá, Rafael Lajara Vizcaíno & Jorge Alberola Lluch22. The Temperature Dependant Efficiency of Photovoltaic Modules - a long termevaluation of experimental measurements 415Jan Machacek, Zdenek Prochazka and Jiri Drapela23. The use of Switched Reluctance Generator in Wind Energy Applications 447Eleonora Darie, Costin Cepişcă and Emanuel Darie24. Tidal Energy Technologies: Currents, Wave and Offshore Wind Power in theUnited Kingdom, Europe and North America 463T. J. Hammons25. Wind Energy Technology 505R. Mesquita Brandão, J. Beleza Carvalho & F. P. Maciel Barbosa26. Wind Generation Modelling for the Management of Electrical TransmissionSystems 531François Vallée27. Variable speed pumped storage hydropower plants for integration of wind powerin isolated power systems 553Jon Are SuulA Model for Greener Power Generation for North-east Sri Lanka based on Stand-alone Renewable Energy Systems 1A Model for Greener Power Generation for North-east Sri Lanka basedon Stand-alone Renewable Energy SystemsReggie DavidrajuhX A Model for Greener Power Generation for North-east Sri Lanka based on Stand-alone Renewable Energy Systems Reggie Davidrajuh University of Stavanger Norway 1. IntroductionIn Northern and Eastern Sri Lanka (NE-SL), electricity is not available in most of the places due to the civil war. There is practically no electricity production and distribution system available or planned. Since Sri Lanka, as a whole, is suffering from power-shortage, it will be not possible for her to supply electricity to NE-SL, by the time the civil war comes to an end. An alternative generation and distribution system should, therefore, be designed to supply electricity to NE-SL. The need for electric power may be analyzed by a fairly intensive and deep study, which invariably requires life cycle analysis of the society in terms of energy use and conservation, industrial and household development, population and industry distribution, export and import of electricity, etc. In contemporary NE-SL, however, the very basics in a societal infrastructure such as industry, highways and education system, are either non-existent or in primitive form. And due to the on-going war in this region, important data for electricity power sector decision-making (such as income and purchasing power of the population, planned transportation and highway infrastructure, environmentally sensitive areas, meteorological data, etc.) are either not available or inadequate. The objective of this chapter is to demonstrate that to achieve a sustainable electricity generation for NE-SL, in addition to utilizing the abundant intermittent resources NE-SL has, namely solar and wind, biomass production must be given utmost priority. The scope of this chapter is limited to design of a mixture of power plants in the cogeneration system, not involving the other aspects of electricity sector such as transmission and distribution. Limitation of this work: due to the ongoing civil war in the region, there is no recent data available about energy production and consumption in this region; the data used in this chapter is from 1994 to 1996. 1Renewable Energy2 2. Needs and Requirements for Sustainable Power Supply The long-term life cycle scenarios for energy to the planet Earth points to massive utilization of solar power, or its derivatives as hydro-, wind- or biomass power (Weinberg, 1990). Power from tidal waves, geothermal and ocean thermal power is insignificant as far as NE-SL is concerned. Each type of energy has its own economy, life cycle and complications. Direct solar power and wind power is complicated by its diluted form and its variation with time and location. However, some of these variations are predictable and may be used in optimal predictive control of power cogeneration systems (Asbjornsen, 1984). Example of such cogeneration system is a combination of solar, wind and biomass power. The power generating systems interconnected by the power grid is an existing technology. But incorporating intermittent power supplies (like solar, wind) to the grid requires careful planning. Nuclear power or hydropower systems acting as the base-load supplier of the utilities, try to even out any fluctuations or failures in the intermittent power supplies, if the intermittent power supplies are connected to the grid. With absence of fossil fuels, hydropower or nuclear power, a sustainable cogeneration system for NE-SL has to use biomass as buffer in incorporating intermittent power supplies. The other option is to include energy storage in the cogeneration system. The energy storage will store energy when the generated electricity from intermittent supply is in excess and it will contribute electricity to the grid when demand exceeds supply. However, studies show that energy storage systems are generally expensive (Friberg, 1993). 2.1 Basic Needs Analysis The total system needs may be formulated as a combined need for electric power by the society and a need for global environmental protection, formulated as follows:  There is a need for electric power, fairly distributed, to the society population  There is a need to replace fossil fuel as energy source for industrial and domestic electric power generation on economic and environmental perspective. 2.2 Basic Requirements Definition The system needs lead to requirements, which are consequences partly of the needs, partly of the customer and user situation, and partly of conceptual solutions. The definition of system requirements is an iterative process, which expands in detail, as the system baseline concept is designed. At the present level of details, and the present system analysis, a set of requirements may be defined fairly simply:  The cost of electric power generation and distribution in NE-SL shall be within the price range for customers and users in the rest Sri Lanka  Compared to the rest of Sri Lanka, the power generation system designed for NE-SL shall have the same or better, reliability, availability and fair distribution of electric power. Cogeneration system should also meet the customers’ power requirement at any time  The power generation system shall have none or minimal harmful pollution effects, on the air, soil or water environment, or on the social environment. There are some other requirements too, due to the varying nature of the intermittent energy sources, and due to the economic requirement on conversion technology:  In case of intermittent primary energy source is used, energy back-up shall be provided by the system  Existing technology for energy conversion shall be utilized to a degree, which is the most feasible economically  The economy of scale shall be utilized optimally in terms of size and capacity of units, modularization and standardization, distribution and centralization 3. System Integration Some thoughts are given below on integrating generators into the electric grid, by considering the technology's social and environmental effects. 3.1 System Response to Customers’ Power Requirements Changes in customers’ power requirements will automatically be taken care of by the demand management system through control of the turbines in the power generation systems. The optimal strategy will be to let the solar power generation run at its maximum capacity at any time. This capacity will change during the day and go to zero when the sun sets. It is also optimal to let the wind power run at its maximum capacity, which will also change with the intensity of the wind. The remaining power requirement will be taken care of by the biomass and fossil fuel power generation systems. As long as the system is designed to tackle all possible situations, as described, the customer will see the power generation system as totally reliable. The key issue in the whole system is to match capacities of the total cogeneration system to the customers’ power requirements and to the availability of the energy sources, solar radiation, and wind speed. Because of the intermittent nature of solar and wind power, when incorporating these generators into the grid, the total plant capacity must always exceed the maximum expected demand by a large margin (penalty for intermittence), in-order to increase system reliability. In NE-SL, where wind and solar power will be contributing most of the time, thermal plants that have higher operating fuel costs but cheaper to build become more attractive (such as natural gas plant), because the reduced operating time will make fuel costs less important. There is a limitation on the extent to which the intermittent sources like solar and wind contribute to the total power generation. As the percentage of power generation of wind and solar increases, there is steady decline in value, because adding intermittent sources reduce the reliability. Contribution from wind energy ranging from 0% to 50% of overall installed capacity is feasible before operational losses become prohibitive (Grubb and Meyer, 1994). 3.2 Smaller the Better For a developing country like NE-SL, which does not have any large-scale hydrologic resource for electricity generation, the selection of optimum power plant mix should be based on small, affordable power generators. Reliability of the power generation and distribution system will be increased if large number of small generators, scattered throughout the nation, is used rather than few large plants. By using a large number of smaller plants, the potential danger of over building or commissioning less cost effective large plants could be avoided. Small plants can be added quickly as they are needed and A Model for Greener Power Generation for North-east Sri Lanka based on Stand-alone Renewable Energy Systems 3Renewable Energy2 2. Needs and Requirements for Sustainable Power Supply The long-term life cycle scenarios for energy to the planet Earth points to massive utilization of solar power, or its derivatives as hydro-, wind- or biomass power (Weinberg, 1990). Power from tidal waves, geothermal and ocean thermal power is insignificant as far as NE-SL is concerned. Each type of energy has its own economy, life cycle and complications. Direct solar power and wind power is complicated by its diluted form and its variation with time and location. However, some of these variations are predictable and may be used in optimal predictive control of power cogeneration systems (Asbjornsen, 1984). Example of such cogeneration system is a combination of solar, wind and biomass power. The power generating systems interconnected by the power grid is an existing technology. But incorporating intermittent power supplies (like solar, wind) to the grid requires careful planning. Nuclear power or hydropower systems acting as the base-load supplier of the utilities, try to even out any fluctuations or failures in the intermittent power supplies, if the intermittent power supplies are connected to the grid. With absence of fossil fuels, hydropower or nuclear power, a sustainable cogeneration system for NE-SL has to use biomass as buffer in incorporating intermittent power supplies. The other option is to include energy storage in the cogeneration system. The energy storage will store energy when the generated electricity from intermittent supply is in excess and it will contribute electricity to the grid when demand exceeds supply. However, studies show that energy storage systems are generally expensive (Friberg, 1993). 2.1 Basic Needs Analysis The total system needs may be formulated as a combined need for electric power by the society and a need for global environmental protection, formulated as follows:  There is a need for electric power, fairly distributed, to the society population  There is a need to replace fossil fuel as energy source for industrial and domestic electric power generation on economic and environmental perspective. 2.2 Basic Requirements Definition The system needs lead to requirements, which are consequences partly of the needs, partly of the customer and user situation, and partly of conceptual solutions. The definition of system requirements is an iterative process, which expands in detail, as the system baseline concept is designed. At the present level of details, and the present system analysis, a set of requirements may be defined fairly simply:  The cost of electric power generation and distribution in NE-SL shall be within the price range for customers and users in the rest Sri Lanka  Compared to the rest of Sri Lanka, the power generation system designed for NE-SL shall have the same or better, reliability, availability and fair distribution of electric power. Cogeneration system should also meet the customers’ power requirement at any time  The power generation system shall have none or minimal harmful pollution effects, on the air, soil or water environment, or on the social environment. There are some other requirements too, due to the varying nature of the intermittent energy sources, and due to the economic requirement on conversion technology:  In case of intermittent primary energy source is used, energy back-up shall be provided by the system  Existing technology for energy conversion shall be utilized to a degree, which is the most feasible economically  The economy of scale shall be utilized optimally in terms of size and capacity of units, modularization and standardization, distribution and centralization 3. System Integration Some thoughts are given below on integrating generators into the electric grid, by considering the technology's social and environmental effects. 3.1 System Response to Customers’ Power Requirements Changes in customers’ power requirements will automatically be taken care of by the demand management system through control of the turbines in the power generation systems. The optimal strategy will be to let the solar power generation run at its maximum capacity at any time. This capacity will change during the day and go to zero when the sun sets. It is also optimal to let the wind power run at its maximum capacity, which will also change with the intensity of the wind. The remaining power requirement will be taken care of by the biomass and fossil fuel power generation systems. As long as the system is designed to tackle all possible situations, as described, the customer will see the power generation system as totally reliable. The key issue in the whole system is to match capacities of the total cogeneration system to the customers’ power requirements and to the availability of the energy sources, solar radiation, and wind speed. Because of the intermittent nature of solar and wind power, when incorporating these generators into the grid, the total plant capacity must always exceed the maximum expected demand by a large margin (penalty for intermittence), in-order to increase system reliability. In NE-SL, where wind and solar power will be contributing most of the time, thermal plants that have higher operating fuel costs but cheaper to build become more attractive (such as natural gas plant), because the reduced operating time will make fuel costs less important. There is a limitation on the extent to which the intermittent sources like solar and wind contribute to the total power generation. As the percentage of power generation of wind and solar increases, there is steady decline in value, because adding intermittent sources reduce the reliability. Contribution from wind energy ranging from 0% to 50% of overall installed capacity is feasible before operational losses become prohibitive (Grubb and Meyer, 1994). 3.2 Smaller the Better For a developing country like NE-SL, which does not have any large-scale hydrologic resource for electricity generation, the selection of optimum power plant mix should be based on small, affordable power generators. Reliability of the power generation and distribution system will be increased if large number of small generators, scattered throughout the nation, is used rather than few large plants. By using a large number of smaller plants, the potential danger of over building or commissioning less cost effective large plants could be avoided. Small plants can be added quickly as they are needed and A Model for Greener Power Generation for North-east Sri Lanka based on Stand-alone Renewable Energy Systems 3Renewable Energy4 even disassembled and moved if loads decline; the economics of the larger plants depend heavily on long-term forecasts (Johannson et al, 1993; Kelly and Weinberg, 1994). In case of biomass-generated electricity, power plants are modest in scale (less than 100MW) to assist collection and transportation of dispersed biomass supplies (Williams and Larson, 1994). If the small generators are scattered throughout the region, a number of other benefits emerges: heavy pollution of a region (congestion) is avoided; employment (less rural exodus) and community participation (awareness) is also increased in distributed power generation. Ideally, each region should become self-sufficient in electricity generation; if the demand exceeds its supply, the utility in that region imports electricity from the neighboring utilities, and similarly, the excess capacity is also sold to the neighboring utilities (Johannson et al, 1993; Kelly and Weinberg, 1994). 3.3 Participatory Development For a sustainable cogeneration system to succeed, the participation of the community becomes essential. Take for example, production of electricity from residues form paddy: today rice husk is normally sold at the rice mills, for insignificant price for trivial use. Now, the rice mill is expected to sell the husk to the utility. The mill management will certainly hesitate to sell if high expenses are incurred in doing so (transportation). However the question is, even if the extra expenses incurred are negligible, whether the management will be willing to sell. The answer to this question depends on the participation of the management in the community’s affairs. This ‘bottom up’ approach where grass roots organize, assist and take responsibilities, and solve (save energy) the energy problem will certainly promote the efficient supply of electricity to the whole nation. 3.4 Household Energy Situation The primary energy consumed in Sri Lanka consists mainly of (Vijayanathan and Lamasena, 1997):  Bio-mass fuel (67%),  Petroleum (22%), and  Hydro-electricity (11%). With the absence fossil fuel resources, biomass is used for both cooking and for small industries, and hence biomass is of vital importance to the economy. Biomass is used by 94% of households use for cooking (Vijayanathan and Lamasena, 1997). Table 1 given below depicts household energy resources for the whole Sri Lanka. Though data for NE-SL is not available, one can safely assume that both fuelwood and kerosene are the crucial energy resources for cooking. Figure 1 show parts of the world where basic remedies like dung and wood are still used for domestic cooking. Type of Energy % Fuelwood and other biomass 87.5 Electricity 6.6 Kerosene 4.8 LAG(LPG) 1.1 Table 1. Biomass resources used for cooking by type Fig. 1. Use of dung and crop residues in developing countries (source: Vijayanathan and Lamasena, 1997) 4. Estimating Electricity Demands of NE-SL by 2010 Since no energy-macroeconomics data for NE-SL is available, the basic assumption made here is that, the energy needs of her population will be the same as the needs of the rest of Sri Lanka. Sri Lanka, with a population of 18 million, today has a per capita electricity consumption of 168 kWh (CIA, 1996). The electricity needs of the NE-SL with a population of 3 million, is therefore 504 GWh per year. In recent years, total firewood consumption in Sri Lanka has been around 5 million tons per year. Of this value, industries consumed approx. 1.05 million tons (Hall et al, 1994). It is reported that the only half of the firewood supply is obtained in a sustainable way (such as forest regeneration, rubber re-plantation cycle etc) (Munasinghe and Meier, 1993). Therefore decline in the natural forest area in Sri Lanka is inevitable. The natural forest area has fallen A Model for Greener Power Generation for North-east Sri Lanka based on Stand-alone Renewable Energy Systems 5Renewable Energy4 even disassembled and moved if loads decline; the economics of the larger plants depend heavily on long-term forecasts (Johannson et al, 1993; Kelly and Weinberg, 1994). In case of biomass-generated electricity, power plants are modest in scale (less than 100MW) to assist collection and transportation of dispersed biomass supplies (Williams and Larson, 1994). If the small generators are scattered throughout the region, a number of other benefits emerges: heavy pollution of a region (congestion) is avoided; employment (less rural exodus) and community participation (awareness) is also increased in distributed power generation. Ideally, each region should become self-sufficient in electricity generation; if the demand exceeds its supply, the utility in that region imports electricity from the neighboring utilities, and similarly, the excess capacity is also sold to the neighboring utilities (Johannson et al, 1993; Kelly and Weinberg, 1994). 3.3 Participatory Development For a sustainable cogeneration system to succeed, the participation of the community becomes essential. Take for example, production of electricity from residues form paddy: today rice husk is normally sold at the rice mills, for insignificant price for trivial use. Now, the rice mill is expected to sell the husk to the utility. The mill management will certainly hesitate to sell if high expenses are incurred in doing so (transportation). However the question is, even if the extra expenses incurred are negligible, whether the management will be willing to sell. The answer to this question depends on the participation of the management in the community’s affairs. This ‘bottom up’ approach where grass roots organize, assist and take responsibilities, and solve (save energy) the energy problem will certainly promote the efficient supply of electricity to the whole nation. 3.4 Household Energy Situation The primary energy consumed in Sri Lanka consists mainly of (Vijayanathan and Lamasena, 1997):  Bio-mass fuel (67%),  Petroleum (22%), and  Hydro-electricity (11%). With the absence fossil fuel resources, biomass is used for both cooking and for small industries, and hence biomass is of vital importance to the economy. Biomass is used by 94% of households use for cooking (Vijayanathan and Lamasena, 1997). Table 1 given below depicts household energy resources for the whole Sri Lanka. Though data for NE-SL is not available, one can safely assume that both fuelwood and kerosene are the crucial energy resources for cooking. Figure 1 show parts of the world where basic remedies like dung and wood are still used for domestic cooking. Type of Energy % Fuelwood and other biomass 87.5 Electricity 6.6 Kerosene 4.8 LAG(LPG) 1.1 Table 1. Biomass resources used for cooking by type Fig. 1. Use of dung and crop residues in developing countries (source: Vijayanathan and Lamasena, 1997) 4. Estimating Electricity Demands of NE-SL by 2010 Since no energy-macroeconomics data for NE-SL is available, the basic assumption made here is that, the energy needs of her population will be the same as the needs of the rest of Sri Lanka. Sri Lanka, with a population of 18 million, today has a per capita electricity consumption of 168 kWh (CIA, 1996). The electricity needs of the NE-SL with a population of 3 million, is therefore 504 GWh per year. In recent years, total firewood consumption in Sri Lanka has been around 5 million tons per year. Of this value, industries consumed approx. 1.05 million tons (Hall et al, 1994). It is reported that the only half of the firewood supply is obtained in a sustainable way (such as forest regeneration, rubber re-plantation cycle etc) (Munasinghe and Meier, 1993). Therefore decline in the natural forest area in Sri Lanka is inevitable. The natural forest area has fallen A Model for Greener Power Generation for North-east Sri Lanka based on Stand-alone Renewable Energy Systems 5Renewable Energy6 to 24% in 1980 from 44% in 1955. Surely NE-SL wants to keep its very limited natural forests therefore the amount of energy obtained in Sri Lanka through firewood supply has to be compensated for through electricity supply. Since firewood is generally the most economic biomass fuel for industrial heating (if used by efficient combustion), only the portion of firewood (ca. 4 million tons) consumed for domestic purposes has to be substituted by electricity. For NE-SL this is equivalent to 0.7 million tons. Therefore additional electricity needed to compensate for use of firewood is 355 GWh. In 1986, Sri Lanka consumed 154.2 thousand tons (kt) of Kerosene, which is solely used for domestic lighting and cooking and 129.0 kt of Fuel oil for industrial use and electricity generation. A portion of fuel oil used for electricity generation is assumed to be 50%. This portion of fuel oil and the amount of kerosene used (totaling 218.7 kt) could be substituted by electricity. Assuming again consumption is proportional to population, this will be equivalent to 130 GWh. The total electricity needed today is therefore approximately 989 GWh. By 2010, the energy needs of the population are expected to be three folded (assuming 8% rate of increase in the standard of living (ca. 6.9%) and population growth (ca. 1.2%)) (CIA, 1996). The total electricity needs per year will then be 2967 GWh. With these data, general performance requirements for the cogeneration system could be set up as shown on Table 2. Functional Requirement The cogeneration system should be able to supply 2967 GWh of electricity per year. Operational Requirement The cogeneration system should be able to support quickly varying demands; should be able to deliver peak power of 564 MW. Physical Requirement The system should be smaller, distributed throughout the nation. Building materials and construction, operation and scrapping conform to international and local standards. Table 2. General requirements of the cogeneration system for the year 2010 5. Energy Resources in NE-SL By 2010, relevant power generation technologies to NE-SL will be solar, wind, biomass, coal and natural gas. 5.1 Biomass Energy from biomass is divided into two general categories 1) agricultural crops and residue, wood and wood waste, animal waste, aquatic plants and 2) plantation biomass: Crop residues: Sugarcane cultivation is a well-established industry in NE-SL. In 1995, NE-SL produced 1.2 million tons of sugarcane (FAOSTAT, 1995). Each ton of sugarcane produce about 0.43 ton of bagasse and leaves. BIG/ISTIG power plant produces 677 kWh of electricity per ton bagasse (Williams and Larson, 1994). If production is assumed to grow at 3.3% (corresponding to Sri Lankas current growth in agricultural production) and all the residues from the sugar industry are used for electricity production, they could produce 568 GWh of electricity by 2010. This is about 19% of the anticipated electricity needs of NE-SL. Residues from the pulp industry, sawdust and bark from de-barking and chipping of logs, could also be used as biomass energy source to the power plant; It is therefore possible for a collateral generation of electricity - pulp - sugar production, especially suited for Eastern Sri Lanka. In NE-SL, agriculture is and will be the main source of income. In 1995, Sri Lanka produced 2.2 million tons of rice (FAOSTAT, 1995). Before the civil war broke out, NE-SL used to produce about 33% of the total rice production in Sri Lanka (TEEDOR, 1992). Each ton of rice produced is accompanied by 2.25 tons of residues in terms of paddy straws and rice husks. With the same rate of growth in production (3.3%), and assuming NE-SL will be able to produce 33% of total production, the residues will have an energy equivalent of 900 GWh of electricity which is about 30% of the electricity needed by year 2010 (paddy rice is a C3 plant, which has about 50% energy contents of a C4 plant such as sugarcane). In NE-SL, Groundnut, Maize, Soybean, Cowpea are also cultivated. Substantial amounts of energy can be generated from residues of these crops. In summary, the large amounts of crop residues that are normally not fully utilized could be used to generate more than 49% of electricity needs. Plantation biomass: One of the tropical legumes, Leucaena leucocephala ("Ipil-Ipil") has already been successfully planted in Sri Lanka. In NE-SL, experimental Casuarina plantation gave promising results (TEEDOR, 1994; TEEDOR, 1995). Plantation biomass energy is a solid option for electricity production in NE-SL. For plantation biomass to be successful plantation wood should be cheaper than the wood purchased on the open market and its cost of energy should be lower than that of conventional energy systems. Also, energy crops should not have long-term environmental, ecological damages. So far no serious study has been done on the economics of harnessing energy crops in Sri Lanka or NE-SL. 5.2 Solar Energy The utilization of solar energy (Photovoltaic or solar thermal) is a major option to satisfy the energy needs of NE-SL. NE-SL, being a tropical country receives annual average solar insulation of 200 W/m2. The solar power system is renewable, it answers all the serious global as well as local environmental issues such as, zero emission of Green House Gases (GHG) and no solid waste. Already a number of micro projects are being operated in many parts of NE-SL, funded by Non-governmental organizations (NGOs) like TECH-German (Tech, 2009a). Figures 2 shows a solar micro power plant in operation to help education of the youth in most rural areas of NE-SL; In order to improve the education of the rural students, NGOs like TECH-German provide solar panels to schools, libraries, community centers where the students can gather at night and study. A Model for Greener Power Generation for North-east Sri Lanka based on Stand-alone Renewable Energy Systems 7Renewable Energy6 to 24% in 1980 from 44% in 1955. Surely NE-SL wants to keep its very limited natural forests therefore the amount of energy obtained in Sri Lanka through firewood supply has to be compensated for through electricity supply. Since firewood is generally the most economic biomass fuel for industrial heating (if used by efficient combustion), only the portion of firewood (ca. 4 million tons) consumed for domestic purposes has to be substituted by electricity. For NE-SL this is equivalent to 0.7 million tons. Therefore additional electricity needed to compensate for use of firewood is 355 GWh. In 1986, Sri Lanka consumed 154.2 thousand tons (kt) of Kerosene, which is solely used for domestic lighting and cooking and 129.0 kt of Fuel oil for industrial use and electricity generation. A portion of fuel oil used for electricity generation is assumed to be 50%. This portion of fuel oil and the amount of kerosene used (totaling 218.7 kt) could be substituted by electricity. Assuming again consumption is proportional to population, this will be equivalent to 130 GWh. The total electricity needed today is therefore approximately 989 GWh. By 2010, the energy needs of the population are expected to be three folded (assuming 8% rate of increase in the standard of living (ca. 6.9%) and population growth (ca. 1.2%)) (CIA, 1996). The total electricity needs per year will then be 2967 GWh. With these data, general performance requirements for the cogeneration system could be set up as shown on Table 2. Functional Requirement The cogeneration system should be able to supply 2967 GWh of electricity per year. Operational Requirement The cogeneration system should be able to support quickly varying demands; should be able to deliver peak power of 564 MW. Physical Requirement The system should be smaller, distributed throughout the nation. Building materials and construction, operation and scrapping conform to international and local standards. Table 2. General requirements of the cogeneration system for the year 2010 5. Energy Resources in NE-SL By 2010, relevant power generation technologies to NE-SL will be solar, wind, biomass, coal and natural gas. 5.1 Biomass Energy from biomass is divided into two general categories 1) agricultural crops and residue, wood and wood waste, animal waste, aquatic plants and 2) plantation biomass: Crop residues: Sugarcane cultivation is a well-established industry in NE-SL. In 1995, NE-SL produced 1.2 million tons of sugarcane (FAOSTAT, 1995). Each ton of sugarcane produce about 0.43 ton of bagasse and leaves. BIG/ISTIG power plant produces 677 kWh of electricity per ton bagasse (Williams and Larson, 1994). If production is assumed to grow at 3.3% (corresponding to Sri Lankas current growth in agricultural production) and all the residues from the sugar industry are used for electricity production, they could produce 568 GWh of electricity by 2010. This is about 19% of the anticipated electricity needs of NE-SL. Residues from the pulp industry, sawdust and bark from de-barking and chipping of logs, could also be used as biomass energy source to the power plant; It is therefore possible for a collateral generation of electricity - pulp - sugar production, especially suited for Eastern Sri Lanka. In NE-SL, agriculture is and will be the main source of income. In 1995, Sri Lanka produced 2.2 million tons of rice (FAOSTAT, 1995). Before the civil war broke out, NE-SL used to produce about 33% of the total rice production in Sri Lanka (TEEDOR, 1992). Each ton of rice produced is accompanied by 2.25 tons of residues in terms of paddy straws and rice husks. With the same rate of growth in production (3.3%), and assuming NE-SL will be able to produce 33% of total production, the residues will have an energy equivalent of 900 GWh of electricity which is about 30% of the electricity needed by year 2010 (paddy rice is a C3 plant, which has about 50% energy contents of a C4 plant such as sugarcane). In NE-SL, Groundnut, Maize, Soybean, Cowpea are also cultivated. Substantial amounts of energy can be generated from residues of these crops. In summary, the large amounts of crop residues that are normally not fully utilized could be used to generate more than 49% of electricity needs. Plantation biomass: One of the tropical legumes, Leucaena leucocephala ("Ipil-Ipil") has already been successfully planted in Sri Lanka. In NE-SL, experimental Casuarina plantation gave promising results (TEEDOR, 1994; TEEDOR, 1995). Plantation biomass energy is a solid option for electricity production in NE-SL. For plantation biomass to be successful plantation wood should be cheaper than the wood purchased on the open market and its cost of energy should be lower than that of conventional energy systems. Also, energy crops should not have long-term environmental, ecological damages. So far no serious study has been done on the economics of harnessing energy crops in Sri Lanka or NE-SL. 5.2 Solar Energy The utilization of solar energy (Photovoltaic or solar thermal) is a major option to satisfy the energy needs of NE-SL. NE-SL, being a tropical country receives annual average solar insulation of 200 W/m2. The solar power system is renewable, it answers all the serious global as well as local environmental issues such as, zero emission of Green House Gases (GHG) and no solid waste. Already a number of micro projects are being operated in many parts of NE-SL, funded by Non-governmental organizations (NGOs) like TECH-German (Tech, 2009a). Figures 2 shows a solar micro power plant in operation to help education of the youth in most rural areas of NE-SL; In order to improve the education of the rural students, NGOs like TECH-German provide solar panels to schools, libraries, community centers where the students can gather at night and study. A Model for Greener Power Generation for North-east Sri Lanka based on Stand-alone Renewable Energy Systems 7Renewable Energy8 Fig. 2. Solar panel installed in NE-SL 5.3 Wind Energy Wind power generation offers many benefits too such as, environmentally friendly (no gaseous emissions), rapid modular addition of capacity, wide range of capacities (from few kW to hundreds of MW), easy to integrate with existing power grid, coexistence with other use of land, no fuel needed and economically competitive (NRC, 1991). NE-SL confines to a region where the wind energy is attractive for power generation with average wind speeds over 5 m/s, giving an average generation of over 33% of rated power (Twidell and Weir, 1986). Figures 3 and 4 show a wind power plant in operation at the integrated model farm in Vaddakatchchi, Kilinochchi. Figure 3 shows the wind turbine and figure 4 shows the battery storage system. The 6KW wind mill powers a Dairy Processing unit and a computer center (TECH, 2009b). Fig. 3. Wind Turbine in Vaddakatchchi, Kilinochchi Fig. 4. Battery Storage System A Model for Greener Power Generation for North-east Sri Lanka based on Stand-alone Renewable Energy Systems 9Renewable Energy8 Fig. 2. Solar panel installed in NE-SL 5.3 Wind Energy Wind power generation offers many benefits too such as, environmentally friendly (no gaseous emissions), rapid modular addition of capacity, wide range of capacities (from few kW to hundreds of MW), easy to integrate with existing power grid, coexistence with other use of land, no fuel needed and economically competitive (NRC, 1991). NE-SL confines to a region where the wind energy is attractive for power generation with average wind speeds over 5 m/s, giving an average generation of over 33% of rated power (Twidell and Weir, 1986). Figures 3 and 4 show a wind power plant in operation at the integrated model farm in Vaddakatchchi, Kilinochchi. Figure 3 shows the wind turbine and figure 4 shows the battery storage system. The 6KW wind mill powers a Dairy Processing unit and a computer center (TECH, 2009b). Fig. 3. Wind Turbine in Vaddakatchchi, Kilinochchi Fig. 4. Battery Storage System A Model for Greener Power Generation for North-east Sri Lanka based on Stand-alone Renewable Energy Systems 9Renewable Energy10 6. Estimating the Costs of Energy Production There are a huge number of parameters involved in the equations for calculating costs of energy production. In this section, only a very limited set of parameters is considered. However, the parameters taken into consideration are the most important ones, thus the model developed in this section and in the following section 7, can be considered as representative. 6.1 Estimated Energy Costs in the Year 2010 Fixed annual cost (FAC) of the total installed cost: CRFICCFAC  (1) Where ICC is the total installed capital cost and CRF is the capital recovery factor. CRF is equal to:   nrr 11 (2) Where r represents the interest rate and n is the lifetime in years. The cost of electricity (COE) is then,  8766 CFPOMCFACCOEr$/kWh (3) Where Pr is the rated capacity, OMC is the Operational and Maintenance (O&M) costs per year, and CF is the average capacity factor and 8,766 is the number of hours per year (365.25  24). Table 3 presents values for the parameters involved in the calculations for costs of energy production. All values are given in 1994-US$. The capital recovery factor (0.0888) is for 8% interest rate and a 30-year plant life. All taxes and insurance costs are neglected. Table 4 summarizes the electricity generation costs for different technologies: Biomass (Heaps and Hill, 1996) Solar (De Laquil, 1994) Wind (Cavallo et al, 1994) Coal (Johannson, 1994) System rating (MW) 111 200 500 109 Installed capital cost ($ /kW) 974 3200 850 1122 Variable O&M costs ($/kWh) 0.0009 0.0065 0.008 0.0010 Fixed O&M costs ($ /kW-year) 34.2 - - 48.0 Fuel costs ($ /GJ) 2.43 - - 2.00 CF 0.75 0.59 0.33 0.75 Efficiency 0.43 - - 0.42 Table 3. Parameter values for calculating costs of energy production Wind Biomass Coal Solar Capital (fixed) 2.60 1.32 1.52 5.50 Fuel - 1.18 1.71 - Operation & Maintenance 0.80 0.61 0.83 0.65 Total 3.40 3.11 4.06 6.15 Table 4. Electricity generation costs for different technologies (US cents / kWh). 6.2 Environmental Costs Associated with Electricity Generation It is important to note that the costs of energy given in Table 4 represent only a part of the costs the society has to pay for electricity generation. The costs of electricity generation can be divided into two, namely the direct or internal costs that are borne by the generating company (capital costs and recurrent or operating costs) and the external costs (‘externalities’) such as the costs associated with environmental damage due to electricity production (Hill et al, 1995). Environmental externalities are becoming the most important external costs in electricity generation. There are various economic methods available to internalize (incorporate into the total cost of energy) environmental externalities. In this chapter, internalization values based on US and German studies are used as presented in (Seshadri, 1978). Estimation for air emissions from a coal-fired power plant is given below, in Table 5 (Munasinghe and Meier, 1993). Also given are the values to monetize emissions (Fritsche, 1994). A Model for Greener Power Generation for North-east Sri Lanka based on Stand-alone Renewable Energy Systems 11Renewable Energy10 6. Estimating the Costs of Energy Production There are a huge number of parameters involved in the equations for calculating costs of energy production. In this section, only a very limited set of parameters is considered. However, the parameters taken into consideration are the most important ones, thus the model developed in this section and in the following section 7, can be considered as representative. 6.1 Estimated Energy Costs in the Year 2010 Fixed annual cost (FAC) of the total installed cost: CRFICCFAC  (1) Where ICC is the total installed capital cost and CRF is the capital recovery factor. CRF is equal to:   nrr 11 (2) Where r represents the interest rate and n is the lifetime in years. The cost of electricity (COE) is then,  8766 CFPOMCFACCOEr$/kWh (3) Where Pr is the rated capacity, OMC is the Operational and Maintenance (O&M) costs per year, and CF is the average capacity factor and 8,766 is the number of hours per year (365.25  24). Table 3 presents values for the parameters involved in the calculations for costs of energy production. All values are given in 1994-US$. The capital recovery factor (0.0888) is for 8% interest rate and a 30-year plant life. All taxes and insurance costs are neglected. Table 4 summarizes the electricity generation costs for different technologies: Biomass (Heaps and Hill, 1996) Solar (De Laquil, 1994) Wind (Cavallo et al, 1994) Coal (Johannson, 1994) System rating (MW) 111 200 500 109 Installed capital cost ($ /kW) 974 3200 850 1122 Variable O&M costs ($/kWh) 0.0009 0.0065 0.008 0.0010 Fixed O&M costs ($ /kW-year) 34.2 - - 48.0 Fuel costs ($ /GJ) 2.43 - - 2.00 CF 0.75 0.59 0.33 0.75 Efficiency 0.43 - - 0.42 Table 3. Parameter values for calculating costs of energy production Wind Biomass Coal Solar Capital (fixed) 2.60 1.32 1.52 5.50 Fuel - 1.18 1.71 - Operation & Maintenance 0.80 0.61 0.83 0.65 Total 3.40 3.11 4.06 6.15 Table 4. Electricity generation costs for different technologies (US cents / kWh). 6.2 Environmental Costs Associated with Electricity Generation It is important to note that the costs of energy given in Table 4 represent only a part of the costs the society has to pay for electricity generation. The costs of electricity generation can be divided into two, namely the direct or internal costs that are borne by the generating company (capital costs and recurrent or operating costs) and the external costs (‘externalities’) such as the costs associated with environmental damage due to electricity production (Hill et al, 1995). Environmental externalities are becoming the most important external costs in electricity generation. There are various economic methods available to internalize (incorporate into the total cost of energy) environmental externalities. In this chapter, internalization values based on US and German studies are used as presented in (Seshadri, 1978). Estimation for air emissions from a coal-fired power plant is given below, in Table 5 (Munasinghe and Meier, 1993). Also given are the values to monetize emissions (Fritsche, 1994). A Model for Greener Power Generation for North-east Sri Lanka based on Stand-alone Renewable Energy Systems 11Renewable Energy12 Emission (kg per kWh) Value $/kg Environmental cost $/kWh SO2 0.00640 4.060 0.0260 CO2 0.32000 0.013 0.0042 NOX 0.00370 1.640 0.0061 Particulate 0.00050 2.380 0.0012 HC 0.00004 0.680 0.0000 Solid ashes NA NA NA Limestone NA NA NA Total 0.0375 Table 5. Environmental costs due to coal-fired power generation. 7. Design of Optimal Power Cogeneration Computer modeling and simulations are done to find the optimal mix of power plants in the cogeneration system; most of these computer models are based on linear programming formulation, and the question is to find out "the mix of different types of plants that will satisfy the power demand with minimum costs". The objective function: Minimize S = "The cost of power generation by different plants" subject to constraints: 1. The total generation should meet the demand 2. Each plant’s generation cannot exceed its physical limitation. In mathematical notation (simplified): Minimize:     kk cxS [$ / year] (4) Subject to:  Constraint 1:   kx (5)  Constraint 2: kkk CFXx  (6) Where and , , , , CFXcx  are the generation of a specific plant per year, the Coefficient of Energy (COE) of the specific plant type, the peak power of that plant, and the average capacity factor (CF), respectively. 7.1 Basic Assumptions Two basic assumptions are given below: 1. In order to prevent over-relying on the intermittent solar and wind energy systems, penetration of intermittent power supplies is limited to 33%, if wind power alone is used. If wind and solar power are used (with the assumption that, owing to subsidy solar power costs are now equal to wind power’s) then 40% penetration of these two combined, is chosen. 2. Biomass: Contribution from biomass is taken to be 1468 GWh /year or 170 MW, which is 49% of the total electricity needed. 7.2 Typical Solutions for Cogeneration System Depending on whether solar power is subsided or not, two solutions are proposed:  Case - A In case-A, the environmental damages are neglected. Also, it is assumed that there is no form of subsidy to consider solar power: In this case, a mix of wind farm(s) of 186 MW and biomass power plants totaling to 170 MW could act as the base-load power supplier. Fossil fuel-fired power plants (coal and natural gas) providing 303 MW of power act as the intermediate and peak suppliers. Solar power is too expensive to consider under any category. Total rated power of the plants put together amounts to 659 MW. So there is a reasonable margin (17%) to take care of the intermittent nature of wind energy and also to increase the reliability of the cogeneration system.  Case - B In case-B, the environmental costs are included in the production costs. Also, it is assumed that there are enough subsidies available to consider solar power: In this case, the cogeneration mix becomes: intermittent solar and wind could contribute 226 MW together (40% penetration) to the base load with the addition of 170 MW biomass power plants. Fossil fuel power should provide 281 MW. The ratio of Solar/Wind depends on the subsidy available for solar power generation and meteorological-statistical data. Total rated power of the plants put together amounts to 677 MW. Here too, a reasonable margin (20%) is available to improve the reliability. A Model for Greener Power Generation for North-east Sri Lanka based on Stand-alone Renewable Energy Systems 13Renewable Energy12 Emission (kg per kWh) Value $/kg Environmental cost $/kWh SO2 0.00640 4.060 0.0260 CO2 0.32000 0.013 0.0042 NOX 0.00370 1.640 0.0061 Particulate 0.00050 2.380 0.0012 HC 0.00004 0.680 0.0000 Solid ashes NA NA NA Limestone NA NA NA Total 0.0375 Table 5. Environmental costs due to coal-fired power generation. 7. Design of Optimal Power Cogeneration Computer modeling and simulations are done to find the optimal mix of power plants in the cogeneration system; most of these computer models are based on linear programming formulation, and the question is to find out "the mix of different types of plants that will satisfy the power demand with minimum costs". The objective function: Minimize S = "The cost of power generation by different plants" subject to constraints: 1. The total generation should meet the demand 2. Each plant’s generation cannot exceed its physical limitation. In mathematical notation (simplified): Minimize:     kk cxS [$ / year] (4) Subject to:  Constraint 1:   kx (5)  Constraint 2: kkk CFXx  (6) Where and , , , , CFXcx  are the generation of a specific plant per year, the Coefficient of Energy (COE) of the specific plant type, the peak power of that plant, and the average capacity factor (CF), respectively. 7.1 Basic Assumptions Two basic assumptions are given below: 1. In order to prevent over-relying on the intermittent solar and wind energy systems, penetration of intermittent power supplies is limited to 33%, if wind power alone is used. If wind and solar power are used (with the assumption that, owing to subsidy solar power costs are now equal to wind power’s) then 40% penetration of these two combined, is chosen. 2. Biomass: Contribution from biomass is taken to be 1468 GWh /year or 170 MW, which is 49% of the total electricity needed. 7.2 Typical Solutions for Cogeneration System Depending on whether solar power is subsided or not, two solutions are proposed:  Case - A In case-A, the environmental damages are neglected. Also, it is assumed that there is no form of subsidy to consider solar power: In this case, a mix of wind farm(s) of 186 MW and biomass power plants totaling to 170 MW could act as the base-load power supplier. Fossil fuel-fired power plants (coal and natural gas) providing 303 MW of power act as the intermediate and peak suppliers. Solar power is too expensive to consider under any category. Total rated power of the plants put together amounts to 659 MW. So there is a reasonable margin (17%) to take care of the intermittent nature of wind energy and also to increase the reliability of the cogeneration system.  Case - B In case-B, the environmental costs are included in the production costs. Also, it is assumed that there are enough subsidies available to consider solar power: In this case, the cogeneration mix becomes: intermittent solar and wind could contribute 226 MW together (40% penetration) to the base load with the addition of 170 MW biomass power plants. Fossil fuel power should provide 281 MW. The ratio of Solar/Wind depends on the subsidy available for solar power generation and meteorological-statistical data. Total rated power of the plants put together amounts to 677 MW. Here too, a reasonable margin (20%) is available to improve the reliability. A Model for Greener Power Generation for North-east Sri Lanka based on Stand-alone Renewable Energy Systems 13Renewable Energy14 Case-ACase-Bwind/solarbiomassfossil fuel050100150200250300350Capacity (MW) Fig. 5. Suggestion for an optimal mix of power plants. 0100200300400500600700800Case-A Case-BAnnuitized total installed capital costs (mill US$)fossil fuelbiomasswind/solar Fig. 6. Cost of the cogeneration system. Figures 5 and 6 shows optimal mix of plants in terms of capacity and costs respectively. In Figure 6, coal power replaces fossil fuel power (coal and natural gas) for simplicity. Also, the reduced capital cost of solar power (in Case - B) is assumed to be equal to that of wind power. 8. Conclusion In the typical cogeneration system for NE-SL, the contribution of fossil fuel-fired power is too high: when considering peak rate, it is 53% in case-A, and 50% in case-B (Figure 5). The growing demand for fossil fuel resulting in the rising price of oil, together with the serious environmental damages due to fossil fuel burning indicates that NE-SL should investigate the use of more sustainable resources for her electricity generation. Though intermittent energy resources solar and wind is abundant, their percentage of contribution cannot go beyond certain limit. Therefore the only way out is to engage large-scale plantation biomass energy. In case-B, fossil fuel contributes to 281 MW of electricity generation. This is equivalent to 6.1 million tons of wood. Assuming that one hectare of biomass plantation yields 25 tons of wood, a total area of 109 thousand hectare (5.4% of total area of NE-SL) of biomass plantation is needed for complete sustainable electricity generation. Normally, unproductive or degraded land (unsuitable for field crops) is allocated for plantation thereby avoiding land for food versus fuel conflict. Plantation biomass species like Ipil-Ipil and Casuarina has rapid growth, even on degraded land. In Tamil Nadu, South India, biomass plantation yields an optimal value of 55 dry tons of wood per hectare per year (Seshadri, 1978). Also, Ipil-Ipil and Casuarina requires minimal management and care. Ipil-Ipil has the ability to fix nitrogen from atmosphere and thus has the ‘self-fertilizing’ ability too (a by product of Ipil-Ipil, high-protein seeds and leaves could be used as animal feed and green fertilizer). 9. References Asbornsen, O. (1984) Feed forward predictive and adaptive control by the dynamic matrix. Proceedings of the ACC 1984, pp. 1864-1869 Cavallo, I, Hock, S., and Smith, D. (1994) Wind Energy: Technology and Economics. In Renewable Energy - Sources for Fuels and Electricity, ed. Thomas B. Johansson et. al, Island Press, Washington CIA (1996) World Factbook. CIA, US De Laquil, P. Kearney, D., Geyer, M., and Diver, R. (1994) Solar-thermal Electric Technology. In Renewable Energy - Sources for Fuels and Electricity, edit. Thomas B. Johansson et. al, Island Press, Washington FAOSTAT (1995) Database Food and Agricultural Organisation of the United Nations. FAO, UN Friberg, R. (1993) A photvoltaic solar-hydrogen power plant for rural electrification in India. Part 1. Int. J. Hydrogen Energy, Vol.18, No.10, pp.853-882 (1993) Fritsche, U. (1994) Modelling Externalities: Cost-Effectiveness of Reducing Environmental Impacts. In Renewable Energy - Sources for Fuels and Electricity , edit. Thomas B. Johansson et. al, Island Press, Washington Grubb, M. and Meyer, N. (1994) Wind Energy: Resources systems, and regional strategies. In Renewable Energy - Sources for Fuels and Electricity, edit. Thomas B. Johansson et. al, Island Press, Washington A Model for Greener Power Generation for North-east Sri Lanka based on Stand-alone Renewable Energy Systems 15Annuitized total installed capital costs (mill US$)Renewable Energy14 Case-ACase-Bwind/solarbiomassfossil fuel050100150200250300350Capacity (MW) Fig. 5. Suggestion for an optimal mix of power plants. 0100200300400500600700800Case-A Case-BAnnuitized total installed capital costs (mill US$)fossil fuelbiomasswind/solar Fig. 6. Cost of the cogeneration system. Figures 5 and 6 shows optimal mix of plants in terms of capacity and costs respectively. In Figure 6, coal power replaces fossil fuel power (coal and natural gas) for simplicity. Also, the reduced capital cost of solar power (in Case - B) is assumed to be equal to that of wind power. 8. Conclusion In the typical cogeneration system for NE-SL, the contribution of fossil fuel-fired power is too high: when considering peak rate, it is 53% in case-A, and 50% in case-B (Figure 5). The growing demand for fossil fuel resulting in the rising price of oil, together with the serious environmental damages due to fossil fuel burning indicates that NE-SL should investigate the use of more sustainable resources for her electricity generation. Though intermittent energy resources solar and wind is abundant, their percentage of contribution cannot go beyond certain limit. Therefore the only way out is to engage large-scale plantation biomass energy. In case-B, fossil fuel contributes to 281 MW of electricity generation. This is equivalent to 6.1 million tons of wood. Assuming that one hectare of biomass plantation yields 25 tons of wood, a total area of 109 thousand hectare (5.4% of total area of NE-SL) of biomass plantation is needed for complete sustainable electricity generation. Normally, unproductive or degraded land (unsuitable for field crops) is allocated for plantation thereby avoiding land for food versus fuel conflict. Plantation biomass species like Ipil-Ipil and Casuarina has rapid growth, even on degraded land. In Tamil Nadu, South India, biomass plantation yields an optimal value of 55 dry tons of wood per hectare per year (Seshadri, 1978). Also, Ipil-Ipil and Casuarina requires minimal management and care. Ipil-Ipil has the ability to fix nitrogen from atmosphere and thus has the ‘self-fertilizing’ ability too (a by product of Ipil-Ipil, high-protein seeds and leaves could be used as animal feed and green fertilizer). 9. References Asbornsen, O. (1984) Feed forward predictive and adaptive control by the dynamic matrix. Proceedings of the ACC 1984, pp. 1864-1869 Cavallo, I, Hock, S., and Smith, D. (1994) Wind Energy: Technology and Economics. In Renewable Energy - Sources for Fuels and Electricity, ed. Thomas B. Johansson et. al, Island Press, Washington CIA (1996) World Factbook. CIA, US De Laquil, P. Kearney, D., Geyer, M., and Diver, R. (1994) Solar-thermal Electric Technology. In Renewable Energy - Sources for Fuels and Electricity, edit. Thomas B. Johansson et. al, Island Press, Washington FAOSTAT (1995) Database Food and Agricultural Organisation of the United Nations. FAO, UN Friberg, R. (1993) A photvoltaic solar-hydrogen power plant for rural electrification in India. Part 1. Int. J. Hydrogen Energy, Vol.18, No.10, pp.853-882 (1993) Fritsche, U. (1994) Modelling Externalities: Cost-Effectiveness of Reducing Environmental Impacts. In Renewable Energy - Sources for Fuels and Electricity , edit. Thomas B. Johansson et. al, Island Press, Washington Grubb, M. and Meyer, N. (1994) Wind Energy: Resources systems, and regional strategies. In Renewable Energy - Sources for Fuels and Electricity, edit. Thomas B. Johansson et. al, Island Press, Washington A Model for Greener Power Generation for North-east Sri Lanka based on Stand-alone Renewable Energy Systems 15Annuitized total installed capital costs (mill US$)Renewable Energy16 Hall, D. Rosillo-Calle, F., Williams, R. and Woods, J. (1994) Biomass for energy: supply prospects. In Renewable Energy - Sources for Fuels and Electricity, edit. T. B. Johansson et. al, Island Press, Washington Heaps, C. and Hill, D. (1996) Applying Fuel Chain Analysis in Venezuela and Sri Lanka. In SEI-B Energy & Environment News, Nr.1, Feb 1996 Hill, R., O’Keefe, P. and Snape, C. (1995) The Future of Energy Use, Earthscan Publications Limited, London Johannson, T., Kelly, H., Reddy, A., and Williams, R. (1993) Renewable fuels and electricity for a growing world economy. In Renewable Energy - Sources for Fuels and Electricity, edit. Thomas B. Johansson et. al, Island Press, Washington Kelly, H. and Weinberg, C. (1994) Utility Strategies for using Renewables. In Renewable Energy - Sources for Fuels and Electricity , edit. Thomas B. Johansson et. al, Island Press, Washington Munasinghe, M. and Meier, P. (1993) Energy policy analysis and modeling. Cambridge University Press, Cambridge, UK NRC (1991) Assessment of Research Needs for Wind Turbine Rotor Materials Technology. Committee on Assessment of Research Needs for Wind Turbine Rotor Materials Technology, National Research Council. National Academy Press, Washington, D.C. Seshadri, C. (1978) Energy Plantation - a case study of the Coromandel Littoral. Shri A. M. M. Murugappa Chettiar Research Centre, Madras, India TECH (2009a): Solar Power Micro Power Plant, http://www.techonnet.org/ projects/Project-EP-01.htm TECH (2009b): Wind Power Micro Power Plant, http://www.techonnet.org/projects/ Project-EP-02.htm TEEDOR (1995) Casuarina Plantation - Project description. (In Tamil language). TEEDOR, Thavady, Sri Lanka TEEDOR (1994) Aatharam. Mar-Apr 94. (In Tamil language). TEEDOR, Thavady, Sri Lanka TEEDOR (1992) Food needs and goals to achieve (In Tamil language). TEEDOR, Thavady, Sri Lanka Twidell, J. and Weir, A. (1986) Renewable Energy Sources. E. & F. Spon Ltd., New York Vijayanathan, L., Lamasena, L. (1997): Boiling Point 39: Using biomass residues for energy. Autumn 1997 Weinberg, C. (1990) Energy from the Sun. In: Energy for the Planet Earth. Readings from Scientific American, W. H. Freeman and Co., NY Williams, R. and Larson, E. (1994) Advanced gasification based biomass power generation. In Renewable Energy - Sources for Fuels and Electricity, edit. Thomas B. Johansson et. al, Island Press, Washington Automatic Sun-Tracker System for Photo-Voltaic Plants 17Automatic Sun-Tracker System for Photo-Voltaic PlantsJoao M. G. FigueiredoX Automatic Sun-Tracker System for Photo-Voltaic Plants Joao M. G. Figueiredo CEM-IDMEC, Universidade Évora, Mechatronics Group R. Romão Ramalho, 59; 7000-671 Évora, Portugal Phone/Fax number: +00351 266 745 300, e-mail: jfig@uevora.pt 1. Introduction According to market economy, the increasing worldwide demand for energy, forces a continuous rise on the price of fossil combustibles. In fact, it is expected in the near future, that the demand for energy will grow faster than the finding out of new available fossil resources (Khan et al., 2007). This market behaviour brings a positive challenge to the scientific community as more funds are allocated for the research and development of new alternatives to the usual main energetic sources (fossil combustibles). In this context we have seen, in the last decades, to a concentrated focus on renewable energy research. Among these renewable energetic sources, the international scientific community has devoted intense efforts to wind, solar photovoltaic and biomass. Some investigations and hardware developments on wave energy have been led by Great Britain and Portugal (Wave-Energy-Centre). In this paper an intelligent sun-tracking system for efficiency maximization referring photovoltaic energy production is developed. Nowadays photovoltaic energy has a low efficiency ratio concerning the complete distribution chain from production to consumption (ca. 12%). In optimized environments (materials, electric inverters, tracking systems, etc) an input of 1000W of solar incident energy can bring ca. 190W in electricity (efficiency of 19%). This low performance ratio implies big Earth surface consumption when it is intended to install industrial photovoltaic units with significant production impact (50MW – 100MW). Today it is being built in south Portugal a photovoltaic plant with 64MW production capacity which occupies an huge area of ca. 400 ha (4 Km2). The more relevant side effect of the low efficiency of photovoltaic systems is its poor competition related to traditional combustibles in both economical and financial aspects. Owing to changes in the solar radiation energy and in the cell operating temperature, the output power of a solar array is not constant at all times. Consequently, a maximum solar power tracking controller is always needed in any scheme with solar cell arrays to ensure maximum utilization. Therefore, works to solve the problems on maximum power point (MPP) tracking have always been a hot topic for photovoltaic array utilization systems. A logical MPP tracking search algorithm using normalized current, voltage and power at the 2Renewable Energy16 Hall, D. Rosillo-Calle, F., Williams, R. and Woods, J. (1994) Biomass for energy: supply prospects. In Renewable Energy - Sources for Fuels and Electricity, edit. T. B. Johansson et. al, Island Press, Washington Heaps, C. and Hill, D. (1996) Applying Fuel Chain Analysis in Venezuela and Sri Lanka. In SEI-B Energy & Environment News, Nr.1, Feb 1996 Hill, R., O’Keefe, P. and Snape, C. (1995) The Future of Energy Use, Earthscan Publications Limited, London Johannson, T., Kelly, H., Reddy, A., and Williams, R. (1993) Renewable fuels and electricity for a growing world economy. In Renewable Energy - Sources for Fuels and Electricity, edit. Thomas B. Johansson et. al, Island Press, Washington Kelly, H. and Weinberg, C. (1994) Utility Strategies for using Renewables. In Renewable Energy - Sources for Fuels and Electricity , edit. Thomas B. Johansson et. al, Island Press, Washington Munasinghe, M. and Meier, P. (1993) Energy policy analysis and modeling. Cambridge University Press, Cambridge, UK NRC (1991) Assessment of Research Needs for Wind Turbine Rotor Materials Technology. Committee on Assessment of Research Needs for Wind Turbine Rotor Materials Technology, National Research Council. National Academy Press, Washington, D.C. Seshadri, C. (1978) Energy Plantation - a case study of the Coromandel Littoral. Shri A. M. M. Murugappa Chettiar Research Centre, Madras, India TECH (2009a): Solar Power Micro Power Plant, http://www.techonnet.org/ projects/Project-EP-01.htm TECH (2009b): Wind Power Micro Power Plant, http://www.techonnet.org/projects/ Project-EP-02.htm TEEDOR (1995) Casuarina Plantation - Project description. (In Tamil language). TEEDOR, Thavady, Sri Lanka TEEDOR (1994) Aatharam. Mar-Apr 94. (In Tamil language). TEEDOR, Thavady, Sri Lanka TEEDOR (1992) Food needs and goals to achieve (In Tamil language). TEEDOR, Thavady, Sri Lanka Twidell, J. and Weir, A. (1986) Renewable Energy Sources. E. & F. Spon Ltd., New York Vijayanathan, L., Lamasena, L. (1997): Boiling Point 39: Using biomass residues for energy. Autumn 1997 Weinberg, C. (1990) Energy from the Sun. In: Energy for the Planet Earth. Readings from Scientific American, W. H. Freeman and Co., NY Williams, R. and Larson, E. (1994) Advanced gasification based biomass power generation. In Renewable Energy - Sources for Fuels and Electricity, edit. Thomas B. Johansson et. al, Island Press, Washington Automatic Sun-Tracker System for Photo-Voltaic Plants 17Automatic Sun-Tracker System for Photo-Voltaic PlantsJoao M. G. FigueiredoX Automatic Sun-Tracker System for Photo-Voltaic Plants Joao M. G. Figueiredo CEM-IDMEC, Universidade Évora, Mechatronics Group R. Romão Ramalho, 59; 7000-671 Évora, Portugal Phone/Fax number: +00351 266 745 300, e-mail: jfig@uevora.pt 1. Introduction According to market economy, the increasing worldwide demand for energy, forces a continuous rise on the price of fossil combustibles. In fact, it is expected in the near future, that the demand for energy will grow faster than the finding out of new available fossil resources (Khan et al., 2007). This market behaviour brings a positive challenge to the scientific community as more funds are allocated for the research and development of new alternatives to the usual main energetic sources (fossil combustibles). In this context we have seen, in the last decades, to a concentrated focus on renewable energy research. Among these renewable energetic sources, the international scientific community has devoted intense efforts to wind, solar photovoltaic and biomass. Some investigations and hardware developments on wave energy have been led by Great Britain and Portugal (Wave-Energy-Centre). In this paper an intelligent sun-tracking system for efficiency maximization referring photovoltaic energy production is developed. Nowadays photovoltaic energy has a low efficiency ratio concerning the complete distribution chain from production to consumption (ca. 12%). In optimized environments (materials, electric inverters, tracking systems, etc) an input of 1000W of solar incident energy can bring ca. 190W in electricity (efficiency of 19%). This low performance ratio implies big Earth surface consumption when it is intended to install industrial photovoltaic units with significant production impact (50MW – 100MW). Today it is being built in south Portugal a photovoltaic plant with 64MW production capacity which occupies an huge area of ca. 400 ha (4 Km2). The more relevant side effect of the low efficiency of photovoltaic systems is its poor competition related to traditional combustibles in both economical and financial aspects. Owing to changes in the solar radiation energy and in the cell operating temperature, the output power of a solar array is not constant at all times. Consequently, a maximum solar power tracking controller is always needed in any scheme with solar cell arrays to ensure maximum utilization. Therefore, works to solve the problems on maximum power point (MPP) tracking have always been a hot topic for photovoltaic array utilization systems. A logical MPP tracking search algorithm using normalized current, voltage and power at the 2Renewable Energy18 work points, that corresponds to the maximum power point values for different operating conditions was early tested (Atlas, 1992), (Atlas, 1996). A on-line controller to track the MPPs under changing illumination was described in (Hua & Lin, 2003). An optimization approach using fuzzy was given in (Benlarbi et al., 2004) for PV water pumping systems. Other MPP tracking controllers can be found in (Hua & Lin, 2004) and (Chen et al., 2004). This paper focuses on the optimization of the electric energy production by photovoltaic cells through the development of an intelligent sun-tracking system. The developed tracking system is innovative in relation to the usual sun tracking systems available in the market. The usual available solutions for tracking systems rely on the knowledge of the geographical position of the solar panel on the earth surface. With this knowledge it is possible to know the relative position of the sun, on a time basis, according to the well known solar tables (Solardat). Modern solutions incorporate a GPS system to calculate the position of the solar panel on the Earth surface. The orientations to be followed by the photovoltaic panel, on a regular time-base, are then pre-programmed, on an open loop approach. There are significant efforts on the optimization of sun tracking systems as it is documented by several registered international patents. These solutions are based either on the above described principle either on the quantification of the received solar energy, either on the maximization of the solar incident radiation through the use of light concentration lens or mirrors (Biee & Chace, 2009), (Rubio et al., 2007). The solution developed in this paper is innovative related to the above referred approaches as this system is autonomous regarding the information needed to process the optimal orientation and it is intelligent in a way that it monitors, on a real-time base, the photovoltaic energy production and it avoids systematic failures coming from changes on the assumed blind values (position, initial infrastructure orientation, cleanness of the photovoltaic cells, etc.). 2. System Description 2.1 Overall System Presentation The overall system is presented in fig. 1. The complete strategy is composed by 5 sub-systems: 1) Electro-Mechanical Structure; 2) Control Unit; 3) Supervisory System; 4) Wind-meter; 5) Photovoltaic Park. The developed tracking system searches the optimal orientation of a surface, related to the sun incident radiation. The global performance of the system is described below. The planar surface is composed by a photovoltaic cell which is motorized by 2-orthogonal axis. These two controlled DOF (Degrees Of Freedom) are managed by a PLC (Programmable Logic Controller) according to a search programme that compares the electric power produced by the photovoltaic cell in each correspondent orientation. The maximal power value is stored and the correspondent orientations on both motorized axis are stored. This new optimal orientation of the tracking system is then communicated to the industrial photovoltaic park in order to transfer the new optimal orientation to all PV-production panels. 2.2 Electro-Mechanical Structure The operational subset of the tracking system, named Electro-Mechanical System, is presented in figs 2 and 3. This structure has two DOF, motorized by stepper motors with incorporated encoders, in order to track exactly the prescribed path. The mechanical system was designed using standard industrial Aluminium profiles in order to obtain a simple and economic structure. The mechanical structure is mainly composed by Bosch-Profiles and Aluminium plates. The two motorized axis are composed by Step-motors assembled to Aluminium shafts. Figure 2 illustrates the several main components of the mechatronic system: Part n. 6 = Step-Motor to control axis 1; Part n. 7 = Step-Motor to control axis 2; Part n. 8 = Photovoltaic cell (150mmx150mm). Figure 3 details the two designed degrees of freedom (DOF). 2.3 Control Unit The control unit is composed by a PLC system (Programmable Logic Controller). This control system has the complete operational management of the tracking system. The main tasks performed by the system are: - Control of the two step motors; - Processing the data from both encoders; - Processing the voltage signal coming from the PV-Cell; 123 InternetGSM mobile network54 Fig. 1. Overall System Presentation Automatic Sun-Tracker System for Photo-Voltaic Plants 19Renewable Energy18 work points, that corresponds to the maximum power point values for different operating conditions was early tested (Atlas, 1992), (Atlas, 1996). A on-line controller to track the MPPs under changing illumination was described in (Hua & Lin, 2003). An optimization approach using fuzzy was given in (Benlarbi et al., 2004) for PV water pumping systems. Other MPP tracking controllers can be found in (Hua & Lin, 2004) and (Chen et al., 2004). This paper focuses on the optimization of the electric energy production by photovoltaic cells through the development of an intelligent sun-tracking system. The developed tracking system is innovative in relation to the usual sun tracking systems available in the market. The usual available solutions for tracking systems rely on the knowledge of the geographical position of the solar panel on the earth surface. With this knowledge it is possible to know the relative position of the sun, on a time basis, according to the well known solar tables (Solardat). Modern solutions incorporate a GPS system to calculate the position of the solar panel on the Earth surface. The orientations to be followed by the photovoltaic panel, on a regular time-base, are then pre-programmed, on an open loop approach. There are significant efforts on the optimization of sun tracking systems as it is documented by several registered international patents. These solutions are based either on the above described principle either on the quantification of the received solar energy, either on the maximization of the solar incident radiation through the use of light concentration lens or mirrors (Biee & Chace, 2009), (Rubio et al., 2007). The solution developed in this paper is innovative related to the above referred approaches as this system is autonomous regarding the information needed to process the optimal orientation and it is intelligent in a way that it monitors, on a real-time base, the photovoltaic energy production and it avoids systematic failures coming from changes on the assumed blind values (position, initial infrastructure orientation, cleanness of the photovoltaic cells, etc.). 2. System Description 2.1 Overall System Presentation The overall system is presented in fig. 1. The complete strategy is composed by 5 sub-systems: 1) Electro-Mechanical Structure; 2) Control Unit; 3) Supervisory System; 4) Wind-meter; 5) Photovoltaic Park. The developed tracking system searches the optimal orientation of a surface, related to the sun incident radiation. The global performance of the system is described below. The planar surface is composed by a photovoltaic cell which is motorized by 2-orthogonal axis. These two controlled DOF (Degrees Of Freedom) are managed by a PLC (Programmable Logic Controller) according to a search programme that compares the electric power produced by the photovoltaic cell in each correspondent orientation. The maximal power value is stored and the correspondent orientations on both motorized axis are stored. This new optimal orientation of the tracking system is then communicated to the industrial photovoltaic park in order to transfer the new optimal orientation to all PV-production panels. 2.2 Electro-Mechanical Structure The operational subset of the tracking system, named Electro-Mechanical System, is presented in figs 2 and 3. This structure has two DOF, motorized by stepper motors with incorporated encoders, in order to track exactly the prescribed path. The mechanical system was designed using standard industrial Aluminium profiles in order to obtain a simple and economic structure. The mechanical structure is mainly composed by Bosch-Profiles and Aluminium plates. The two motorized axis are composed by Step-motors assembled to Aluminium shafts. Figure 2 illustrates the several main components of the mechatronic system: Part n. 6 = Step-Motor to control axis 1; Part n. 7 = Step-Motor to control axis 2; Part n. 8 = Photovoltaic cell (150mmx150mm). Figure 3 details the two designed degrees of freedom (DOF). 2.3 Control Unit The control unit is composed by a PLC system (Programmable Logic Controller). This control system has the complete operational management of the tracking system. The main tasks performed by the system are: - Control of the two step motors; - Processing the data from both encoders; - Processing the voltage signal coming from the PV-Cell; 123 InternetGSM mobile network54 Fig. 1. Overall System Presentation Automatic Sun-Tracker System for Photo-Voltaic Plants 19Renewable Energy20 - Processing the data from the external proximity sensors that informs the system about the hard-home position reference. This PLC controls directly the tracking system and commands all other PV-Panels, from the solar Park, through a Profibus-DP network (Siemens, 2001a). Figure 4 shows an example of a solar park with motorized PV-Panels. Figure 5 illustrates the Profibus network implemented in this study. 786 Fig. 2. Electro-Mechanical System: Main Components 21 Fig. 3. Electro-Mechanical System: Axis 1 and 2 2.4 Supervisory System A SCADA system (Supervisory Control And Data Acquisition) is implemented to monitor and supervise the tracking system. A Supervisory Control and Data Acquisition (SCADA) System is used as an application development tool that enables system integrators to create sophisticated supervisory and control applications for a variety of technological domains, mainly in the industry field. The main feature of a SCADA system is its ability to communicate with control equipment in the field, through the PLC network. As the equipment is monitored and data is recorded, a SCADA application responds according to system logic requirements or operator requests. In the developed supervisory system the SCADA application manages the overall system dynamics. The Communication flux between the supervisory system and the control unit is illustrated in fig. 5. The SCADA PC is simultaneously a SCADA server and an internet server, as the implemented SCADA application is web enabled. 3. Experimental Prototype 3.1 Physical Description The prototype built followed the design presented in figure 2. This system incorporates a PV-cell 150mmx150mm, Pmax=1,12W, (Polycrystalline Silicon wafer) and the whole structure is made of aluminium alloy. In fig. 6 the global developed prototype is shown. The Fig. 4. Solar Park – motorized PV panels Automatic Sun-Tracker System for Photo-Voltaic Plants 21Renewable Energy20 - Processing the data from the external proximity sensors that informs the system about the hard-home position reference. This PLC controls directly the tracking system and commands all other PV-Panels, from the solar Park, through a Profibus-DP network (Siemens, 2001a). Figure 4 shows an example of a solar park with motorized PV-Panels. Figure 5 illustrates the Profibus network implemented in this study. 786 Fig. 2. Electro-Mechanical System: Main Components 21 Fig. 3. Electro-Mechanical System: Axis 1 and 2 2.4 Supervisory System A SCADA system (Supervisory Control And Data Acquisition) is implemented to monitor and supervise the tracking system. A Supervisory Control and Data Acquisition (SCADA) System is used as an application development tool that enables system integrators to create sophisticated supervisory and control applications for a variety of technological domains, mainly in the industry field. The main feature of a SCADA system is its ability to communicate with control equipment in the field, through the PLC network. As the equipment is monitored and data is recorded, a SCADA application responds according to system logic requirements or operator requests. In the developed supervisory system the SCADA application manages the overall system dynamics. The Communication flux between the supervisory system and the control unit is illustrated in fig. 5. The SCADA PC is simultaneously a SCADA server and an internet server, as the implemented SCADA application is web enabled. 3. Experimental Prototype 3.1 Physical Description The prototype built followed the design presented in figure 2. This system incorporates a PV-cell 150mmx150mm, Pmax=1,12W, (Polycrystalline Silicon wafer) and the whole structure is made of aluminium alloy. In fig. 6 the global developed prototype is shown. The Fig. 4. Solar Park – motorized PV panels Automatic Sun-Tracker System for Photo-Voltaic Plants 21Renewable Energy22 control unit was developed using an industrial Siemens S7-300 PLC (Programmable Logic Controller). The selected PLC system is a modular device that is constituted by the following modules: Slot1 = Power supply PS 307-2A Slot2 = Processor CPU 315-2DP Slot4 = Communication module CP 342 -5 Slot5 = Digital card DI8/DO8xDC24V/0,5A Slot6 = Analog card AI8 x12bit Slot7 = Analog card AO4 x12bit Slot8 = FM card – Counter Module (FM350) Slot9 = FM card – Counter Module (FM350) Slot10 = FM card – Stepper Motor (FM353) Slot11 = FM card – Stepper Motor (FM353) Additionally, the PLC-tracker has a modem for GSM communication that provides the system capacity to communicate through the mobile phone network. SCADA-PCWorkstationSCADA-PCWorkstationSCADAPCPLC TrackerPV-Panel 1Profibus/ DP(RS 232/ MPI)PV-Panel 2PV-Panel 3PV-Panel 4 Fig. 5. Communication Strategy (SCADA – PLC Tracker) The driving unit is composed by two motorized axis, with the following characteristics: i) Axis 1 - Step motor: Nanotec ST4018L0804, 50Ncm; - Opt. Encoder: HP HEDL-5540 A14, 500 Pulses - Coupling unit: Oldham D5 - Proximity sensor: Omrom EA2 M8 PNP ii) Axis 2 - Step motor: Nanotec ST5918L1008, 170Ncm; - Gear box: Nanotec PLE40-1S-4 - Opt. Encoder: HP HEDL-5540 A14, 500 Pulses - Coupling unit: Oldham D25 - Proximity sensor: Omron EA2 M8 PNP Figure 6 shows a global view of the built Prototype. Fig. 6. Built Prototype: Global view Automatic Sun-Tracker System for Photo-Voltaic Plants 23Renewable Energy22 control unit was developed using an industrial Siemens S7-300 PLC (Programmable Logic Controller). The selected PLC system is a modular device that is constituted by the following modules: Slot1 = Power supply PS 307-2A Slot2 = Processor CPU 315-2DP Slot4 = Communication module CP 342 -5 Slot5 = Digital card DI8/DO8xDC24V/0,5A Slot6 = Analog card AI8 x12bit Slot7 = Analog card AO4 x12bit Slot8 = FM card – Counter Module (FM350) Slot9 = FM card – Counter Module (FM350) Slot10 = FM card – Stepper Motor (FM353) Slot11 = FM card – Stepper Motor (FM353) Additionally, the PLC-tracker has a modem for GSM communication that provides the system capacity to communicate through the mobile phone network. SCADA-PCWorkstationSCADA-PCWorkstationSCADAPCPLC TrackerPV-Panel 1Profibus/ DP(RS 232/ MPI)PV-Panel 2PV-Panel 3PV-Panel 4 Fig. 5. Communication Strategy (SCADA – PLC Tracker) The driving unit is composed by two motorized axis, with the following characteristics: i) Axis 1 - Step motor: Nanotec ST4018L0804, 50Ncm; - Opt. Encoder: HP HEDL-5540 A14, 500 Pulses - Coupling unit: Oldham D5 - Proximity sensor: Omrom EA2 M8 PNP ii) Axis 2 - Step motor: Nanotec ST5918L1008, 170Ncm; - Gear box: Nanotec PLE40-1S-4 - Opt. Encoder: HP HEDL-5540 A14, 500 Pulses - Coupling unit: Oldham D25 - Proximity sensor: Omron EA2 M8 PNP Figure 6 shows a global view of the built Prototype. Fig. 6. Built Prototype: Global view Automatic Sun-Tracker System for Photo-Voltaic Plants 23Renewable Energy24 3.2 Implemented Control Algorithm The software used for the PLC programming was the Siemens Simatic Step 7 (Siemens, 2000). The designed control algorithm was implemented using the Ladder Logic – LAD (Siemens, 2001b). The developed control algorithm is illustrated in fig. 7. Box0: After reset is activated, the system stores the PV-power generated in the actual position, Pactual, in the variable Pin. The system searchs its reference-null position. It moves until it finds the hardhome position (both external proximity sensors on). In this position the system assumes the absolute orientation angles for both axis equal zero (1 = 2 = 0). The maximal Power, Pmax is set to zero. Both counters, C1, C2, are loaded; Box1: After start is activated, the system iniciates the search for the maximal power generated in axis 1, with an angle increment 10. The system stores the power generated in variable P1. Box2: If P1 < Pmax, the system goes to Box 4, and follows for a new position; Box3: If P1 > Pmax, this position is stored in the variables: 1max, 2max. The max. Power value, Pmax is actualized with the new Power value P1; Box4: Counter for axis 1 is updated; Box5: After all orientations for axis 1 are evaluated, regarding a fixed orientation for axis 2, axis 2 is positioned in a new position, with an angle increment 20, and axis 1 returns to its initial position 1=0. The system re-initiates the search for the optimal orientation of axis 1, regarding the new position of axis 2. The information flux returns to box 1. Box6: After all orientations for axis 1 are evaluated, regarding all different positions of axis 2, the system compares the maximal power found (Pmax) with the initial Power generated, before the search process had begun (Pin). If the new Power value is greater than a pre-defined gain, this new correspondent orientation (1max, 2max) is sent to all park panels. If the power gain is not enough, the new found position is not to follow by the other PV-panels. Box7: After a pre-defined time interval (K) the tracker system initiates a new complete search process in both axis. The information flux returns to box 0. reset-Find and positioning the system in Hardhome ref. 1 = 2 = 0;- Pin = Pactual;- Pmax=0;- C1=5; C2=5start-1 = 1 + 10;- Read P1P1OPmax P1 > Pmax-1max = 1;-2max = 2;-Pmax = P1C1 = C1 - 1C1 =1C1 = 0 -1 = 0;-C1 = 5;-C2 = C2 – 1-2 = 2 + 20;C2 =1C2 = 0T = K [s]T1 = 1012 3457Pmax > Gmin x PinYes- Send new orientation (1max, 2max)to solar park;-Ok = 16NoOk = 1 Fig. 7. Control Algorithm for the Tracking System A short description of the tasks performed by the tracker controller, regarding the above referred algorithm, is described together with fig. 7. 3.3 SCADA Supervisory System The Scada system was developed over the platform Siemens WinCC (Siemens, 2005). The SCADA system used to implement this monitoring and control strategy permits the selective access to the application, depending on the user’s responsibility degree. In this paper we developed three user levels: Operators, Supervisors and Administrators. Several SCADA menus were built. The main characteristic of a SCADA Menu is to be simple, explicit and quick on transmitting the information to the operator or to the System administrator. One of the developed Graphical User Interfaces (GUI) is shown in fig 8. As this SCADA platform is web enabled, all the GUI displayed data is also on-line accessible through the internet. In fig. 9 it is shown the developed main menu for the sun-tracker system. The on-line available information, referring actual data from the tracker unit is: actual position for both axis, actual PV-power generated, max. daily PV-power generated, actual efficiency ratio. 4. Conclusion This paper focus the optimization of the electric energy production by photovoltaic cells through the development of an intelligent sun-tracking system. The developed tracking Fig. 8. Sun Tracker System: SCADA main Menu Automatic Sun-Tracker System for Photo-Voltaic Plants 25Renewable Energy24 3.2 Implemented Control Algorithm The software used for the PLC programming was the Siemens Simatic Step 7 (Siemens, 2000). The designed control algorithm was implemented using the Ladder Logic – LAD (Siemens, 2001b). The developed control algorithm is illustrated in fig. 7. Box0: After reset is activated, the system stores the PV-power generated in the actual position, Pactual, in the variable Pin. The system searchs its reference-null position. It moves until it finds the hardhome position (both external proximity sensors on). In this position the system assumes the absolute orientation angles for both axis equal zero (1 = 2 = 0). The maximal Power, Pmax is set to zero. Both counters, C1, C2, are loaded; Box1: After start is activated, the system iniciates the search for the maximal power generated in axis 1, with an angle increment 10. The system stores the power generated in variable P1. Box2: If P1 < Pmax, the system goes to Box 4, and follows for a new position; Box3: If P1 > Pmax, this position is stored in the variables: 1max, 2max. The max. Power value, Pmax is actualized with the new Power value P1; Box4: Counter for axis 1 is updated; Box5: After all orientations for axis 1 are evaluated, regarding a fixed orientation for axis 2, axis 2 is positioned in a new position, with an angle increment 20, and axis 1 returns to its initial position 1=0. The system re-initiates the search for the optimal orientation of axis 1, regarding the new position of axis 2. The information flux returns to box 1. Box6: After all orientations for axis 1 are evaluated, regarding all different positions of axis 2, the system compares the maximal power found (Pmax) with the initial Power generated, before the search process had begun (Pin). If the new Power value is greater than a pre-defined gain, this new correspondent orientation (1max, 2max) is sent to all park panels. If the power gain is not enough, the new found position is not to follow by the other PV-panels. Box7: After a pre-defined time interval (K) the tracker system initiates a new complete search process in both axis. The information flux returns to box 0. reset-Find and positioning the system in Hardhome ref. 1 = 2 = 0;- Pin = Pactual;- Pmax=0;- C1=5; C2=5start-1 = 1 + 10;- Read P1P1OPmax P1 > Pmax-1max = 1;-2max = 2;-Pmax = P1C1 = C1 - 1C1 =1C1 = 0 -1 = 0;-C1 = 5;-C2 = C2 – 1-2 = 2 + 20;C2 =1C2 = 0T = K [s]T1 = 1012 3457Pmax > Gmin x PinYes- Send new orientation (1max, 2max)to solar park;-Ok = 16NoOk = 1 Fig. 7. Control Algorithm for the Tracking System A short description of the tasks performed by the tracker controller, regarding the above referred algorithm, is described together with fig. 7. 3.3 SCADA Supervisory System The Scada system was developed over the platform Siemens WinCC (Siemens, 2005). The SCADA system used to implement this monitoring and control strategy permits the selective access to the application, depending on the user’s responsibility degree. In this paper we developed three user levels: Operators, Supervisors and Administrators. Several SCADA menus were built. The main characteristic of a SCADA Menu is to be simple, explicit and quick on transmitting the information to the operator or to the System administrator. One of the developed Graphical User Interfaces (GUI) is shown in fig 8. As this SCADA platform is web enabled, all the GUI displayed data is also on-line accessible through the internet. In fig. 9 it is shown the developed main menu for the sun-tracker system. The on-line available information, referring actual data from the tracker unit is: actual position for both axis, actual PV-power generated, max. daily PV-power generated, actual efficiency ratio. 4. Conclusion This paper focus the optimization of the electric energy production by photovoltaic cells through the development of an intelligent sun-tracking system. The developed tracking Fig. 8. Sun Tracker System: SCADA main Menu Automatic Sun-Tracker System for Photo-Voltaic Plants 25Renewable Energy26 system is innovative in relation to the usual sun tracking systems available in the market. In fact, the developed solution has many advantages in relation to similar existing devices, as this system is autonomous regarding the information needed to process the optimal orientation and is intelligent in a way that it performs on-line monitoring of the photovoltaic energy production. The increase in power generation, in relation to other PV-systems, without tracking devices, is of similar magnitude (ca. 25%) as for other usual tracking solutions. However, this system has a relative advantage, as it measures exactly the controlled variable: the actual PV-power generation. 5. References Atlas I, Sharaf A. (1992). “A Fuzzy Logic Power Tracking Controller for a Photovoltaic Energy Conversion Scheme”; Electr. Power Syst. Res. J., 1992; 25 (3); pp. 227-238 Atlas I, Sharaf A. (1996); “A Novel on-line MPP search algorithm for PV arrays”; IEEE Trans. Energy Convers., 1996; 11 (4); pp. 748-754 Benlarbi K., Mokrani L., Nait-Said M. (2004), “A Fuzzy Global Efficiency Optimization of a Photovoltaic Water Pumping System”; Sol Energy 2004; 77; pp. 203-216. Biee, S.; Chace JR, A. (2009). “Solar Tracking Reflector System for Structure Lighting”. European Patent Office, IPC: H01L31/052, EC: E04D13/03; US2009084431 (A1); 2009-04-02 Chen Y., Liu Y., Wu F. (2004) “Multi-Input Converter with Power Factor Correction, Maximum Power Point Tracking, and Ripple-free Input Currents”; IEEE Trans. Power Electron. 2004, 19 (3), pp. 631-639 Hua, C., Lin, J. (2003); “An on-line MPPT algorithm for rapidly changing illuminations of Solar arrays”; Renew Energy 2003; 28; pp. 1129-1142. Hua C., Lin J. (2004) “A modified tracking algorithm for maximum power tracking of solar array”; Energy Conversion and Management 2004, Vol. 45, pp. 911-925. Khan, N., Mariun, Z., Saleem, N., Abas, N. (2007). “Fossil Fuels, New Energy Sources and the Great Energy Crisis”. Renewable and Sustainable Energy Rev (2007), doi:10.1016/j.rser.2007.11 .011 http://solardat.uoregon.edu/ SolarPositionCalculatorhtml http://www.wave-energy-centre.org Rubio, F., Ortega, M., Gordillo, F., López-Martínez, M. (2007). “Application of new control strategy for sun tracking”; Energy Conversion and Management, Vol. 48, Issue 7, July 2007, Pages 2174-2184. Siemens (2000). System Software for S7-300 and S7-400 – Reference Manual, SIEMENS 08/2000; A5E00069892-02 Siemens (2001a). Simatic Net – NCM S7 for Profibus/ FMS. SIEMENS 12/2001. Siemens (2001b). Simatic S7-300 – Ladder Logic (LAD) for S7-300, SIEMENS, 2001. Siemens (2005). Simatic WinCC V6.0 SP2, SIEMENS, 2005. Development of Space-Based Solar Power 27Development of Space-Based Solar PowerLyle M. JenkinsX Development of Space-Based Solar Power Lyle M. Jenkins Jenkins Enterprises Houston, Texas Jenkins1166@sbcglobal.net 1. Overview The potential of space-based solar power (SBSP) as a resource to alleviate world energy needs has not been effective in obtaining the political support for a comprehensive program of evaluation and demonstration. An alternative approach is to emphasize the environmental benefits. Not fully understanding the stability of the Earth System and the specific feedback mechanisms controlling our climate, scientists are unable to effectively predict the course of change to the global environment. Geologic records show a potential rate of change that leaves civilization vulnerable to severe economic effects in a period of significant population growth. Solar energy as an alternative to fossil fuel reduces stress on the Earth’s environmental system. Cost of solar power, particularly from space, is not competitive with current prices of fossil fuels. Collecting the energy in space provides significant advantages in continuity of supply over terrestrial solar, but there is large initial cost prior to getting a return on the investment. The “Fresh Look at Space Solar Power” study shows that concepts needing less investment in an operational system may be feasible. Resources are needed to develop technology and to demonstrate practicality. 2. SBSP Concept Definition The SBSP concept is to collect energy from the Sun in Earth orbit. The electrical energy is converted to microwave frequency for transmission to the surface of the Earth. There it is converted back in to electricity for use. Possible usages are base-load power, fuel conversion or direct delivery to consumers in isolated locations. The available potential of solar energy is greater than energy in petroleum reserves. The primary issue is defining the path to development of SBSP capability. No scientific or technological breakthroughs are needed to develop SBSP. Certain technology may require demonstration such as microwave power transmission through the atmosphere. The primary questions relate to beam focus and efficiency. The challenge will to limit the cost of the required hardware. One of the requirements will be a minimum weight for the particular subsystems that make up the orbital system. This is directly related to the cost of launching the total mass into orbit. 3Renewable Energy26 system is innovative in relation to the usual sun tracking systems available in the market. In fact, the developed solution has many advantages in relation to similar existing devices, as this system is autonomous regarding the information needed to process the optimal orientation and is intelligent in a way that it performs on-line monitoring of the photovoltaic energy production. The increase in power generation, in relation to other PV-systems, without tracking devices, is of similar magnitude (ca. 25%) as for other usual tracking solutions. However, this system has a relative advantage, as it measures exactly the controlled variable: the actual PV-power generation. 5. References Atlas I, Sharaf A. (1992). “A Fuzzy Logic Power Tracking Controller for a Photovoltaic Energy Conversion Scheme”; Electr. Power Syst. Res. J., 1992; 25 (3); pp. 227-238 Atlas I, Sharaf A. (1996); “A Novel on-line MPP search algorithm for PV arrays”; IEEE Trans. Energy Convers., 1996; 11 (4); pp. 748-754 Benlarbi K., Mokrani L., Nait-Said M. (2004), “A Fuzzy Global Efficiency Optimization of a Photovoltaic Water Pumping System”; Sol Energy 2004; 77; pp. 203-216. Biee, S.; Chace JR, A. (2009). “Solar Tracking Reflector System for Structure Lighting”. European Patent Office, IPC: H01L31/052, EC: E04D13/03; US2009084431 (A1); 2009-04-02 Chen Y., Liu Y., Wu F. (2004) “Multi-Input Converter with Power Factor Correction, Maximum Power Point Tracking, and Ripple-free Input Currents”; IEEE Trans. Power Electron. 2004, 19 (3), pp. 631-639 Hua, C., Lin, J. (2003); “An on-line MPPT algorithm for rapidly changing illuminations of Solar arrays”; Renew Energy 2003; 28; pp. 1129-1142. Hua C., Lin J. (2004) “A modified tracking algorithm for maximum power tracking of solar array”; Energy Conversion and Management 2004, Vol. 45, pp. 911-925. Khan, N., Mariun, Z., Saleem, N., Abas, N. (2007). “Fossil Fuels, New Energy Sources and the Great Energy Crisis”. Renewable and Sustainable Energy Rev (2007), doi:10.1016/j.rser.2007.11 .011 http://solardat.uoregon.edu/ SolarPositionCalculatorhtml http://www.wave-energy-centre.org Rubio, F., Ortega, M., Gordillo, F., López-Martínez, M. (2007). “Application of new control strategy for sun tracking”; Energy Conversion and Management, Vol. 48, Issue 7, July 2007, Pages 2174-2184. Siemens (2000). System Software for S7-300 and S7-400 – Reference Manual, SIEMENS 08/2000; A5E00069892-02 Siemens (2001a). Simatic Net – NCM S7 for Profibus/ FMS. SIEMENS 12/2001. Siemens (2001b). Simatic S7-300 – Ladder Logic (LAD) for S7-300, SIEMENS, 2001. Siemens (2005). Simatic WinCC V6.0 SP2, SIEMENS, 2005. Development of Space-Based Solar Power 27Development of Space-Based Solar PowerLyle M. JenkinsX Development of Space-Based Solar Power Lyle M. Jenkins Jenkins Enterprises Houston, Texas Jenkins1166@sbcglobal.net 1. Overview The potential of space-based solar power (SBSP) as a resource to alleviate world energy needs has not been effective in obtaining the political support for a comprehensive program of evaluation and demonstration. An alternative approach is to emphasize the environmental benefits. Not fully understanding the stability of the Earth System and the specific feedback mechanisms controlling our climate, scientists are unable to effectively predict the course of change to the global environment. Geologic records show a potential rate of change that leaves civilization vulnerable to severe economic effects in a period of significant population growth. Solar energy as an alternative to fossil fuel reduces stress on the Earth’s environmental system. Cost of solar power, particularly from space, is not competitive with current prices of fossil fuels. Collecting the energy in space provides significant advantages in continuity of supply over terrestrial solar, but there is large initial cost prior to getting a return on the investment. The “Fresh Look at Space Solar Power” study shows that concepts needing less investment in an operational system may be feasible. Resources are needed to develop technology and to demonstrate practicality. 2. SBSP Concept Definition The SBSP concept is to collect energy from the Sun in Earth orbit. The electrical energy is converted to microwave frequency for transmission to the surface of the Earth. There it is converted back in to electricity for use. Possible usages are base-load power, fuel conversion or direct delivery to consumers in isolated locations. The available potential of solar energy is greater than energy in petroleum reserves. The primary issue is defining the path to development of SBSP capability. No scientific or technological breakthroughs are needed to develop SBSP. Certain technology may require demonstration such as microwave power transmission through the atmosphere. The primary questions relate to beam focus and efficiency. The challenge will to limit the cost of the required hardware. One of the requirements will be a minimum weight for the particular subsystems that make up the orbital system. This is directly related to the cost of launching the total mass into orbit. 3Renewable Energy28 2Cheap, reliable access to space is a key issue in making SBSP economically viable. The mass to be deployed will mandate a reusable launch system. Trades of the number of stages will be needed to optimize the efficiency. One evaluation may be the air launch concept being developed by Burt Rutan. It enables the launch of the upper stages above much of the sensible atmosphere. This reduces aerodynamic loads, but may be limited by a reasonable takeoff weight. Current assembly concepts have assumed construction in low earth orbit. After completion, the solar power satellite would be transferred to a higher orbit. Propulsion to accomplish this is a critical issue. One concept that has the specific impulse to make transfer practical is Variable Specific Impulse Magnetoplasma Rocket (VASIMR). NASA spinoff firm, the Ad Astra Rocket Company, has announced a key milestone in ground testing of its prototype plasma drive technology, The VASIMR "helicon first stage" - which generates the plasma for acceleration by the rest of the drive has achieved its full rated power of 30 kilowatts using Argon propellant, according to the company. This paves the way for further trials in which the ion-cyclotron second stage will get to boost the helicon plasma stream to the target power of 200 kW. The idea of the plasma drives is to use electric power to blast reaction mass (in this case Argon) from its rocket nozzles at a much higher speed than regular chemical rockets can achieve. This means that the carrying spacecraft gets a lot more acceleration from a given amount of fuel. A potential demonstration for VASIMR is maintaining the orbit of the international space station (ISS) without the need to burn large amounts of chemical rocket fuel. This serves as a demonstration of the transfer large structures between orbits. Since the solar power satellite was studied in the late 1970’s there have been many advancements in subsystem technology. These advances have included (a) improvements in photo-voltaic efficiency from about 10% (1970s) to more than 40% (2007); (b) increases in robotics capabilities from simple teleoperated manipulators in a few degrees of freedom (1970s) to fully autonomous robotics with insect-class intelligence and 30-100 degrees of freedom (2007); (c) increases in the efficiency of solid-state devices from around 20% (1970s) to as much as 70%-90% (2007); (d) improvements in materials for structures from simple aluminum (1970s) to advanced composites including nanotechnology composites (2007); (e) the application to large space structures; (f) high temperature super-conductors and many other technologies may be integrated into the design. Microwave beams are constant and conversion efficiencies high. They can be beamed at densities substantially lower than that of sunlight. This delivers more energy per area than terrestrial solar energy. The peak density of the beam can be significantly less than noon sunlight, and at the edge of the rectenna equivalent to the leakage of a microwave oven. This low energy density and choice of wavelength also means that biological effects are likely to be low. The safety of wild life wandering into the beam is not expected to be an issue. The physics of electromagnetic energy beaming is uncompromising. The size of the antenna makes microwave beaming unsuitable as a “secret” weapon. The distance from the 3geostationary belt is so great that beams diverge beyond the coherence and power concentration needed for a weapon. The beam is likely to be designed to require a pilot beam transmitted from the rectenna site. Absent the pilot signal, the system can be programmed to go into an incoherent mode. Concerns may also be addressed through an inspection regime. The likelihood of the beam wandering over a city is extremely low. Even if it occurred, it would not be a hazard. Wireless energy transfer by laser beam represents a different set of requirements. To achieve comparable efficiency, the beam must be more intense. The clouds in the atmosphere will reduce the transfer. The intense beam may produce a hazardous level to be avoided by aircraft and satellites. Still for the application to military power supply, it may be a manageable method. At present, the United States has very limited capabilities to build large structures, very large aperture antennas or very high power systems in orbit. The capability to control and maneuver these systems in space must be developed and demonstrated. Presently, the ability to translate large mass between Earth orbits will be required for deployment SBSP. Eventually, the capability for in-space manufacturing and construction or in-situ space resource utilization may be developed, but at this point it is a challenge that should not be incorporated into the program. One critical item to be demonstrated is capability for beamed power and application to propulsion of large space systems. The Thunderstorm Solar Power Satellite (TSPS) is a concept for interacting with thunderstorms to prevent formation of tornadoes. TSPS benefits are saving lives and reducing property. These benefits are not as sensitive to the system economics as the commercial solar power satellite and justify government investment in space solar power. The TSPS can develop and demonstrate the technology and operations critical to understanding the cost of space solar power. Consequently, there is no direct competition with fossil fuel based power supplies until SSP technology and operations have been demonstrated. Before weather modification can be safely attempted, the fine structure of thunderstorms must be simulated and related to tornadogenesis 3. Environmental Benefits Advocates of space solar power have been presenting the concepts as a means to help meet world energy needs. This argument has not been effective in garnering support for even basic research and technology development. Fossil fuel alternatives have been too cheap and near term effect on the “economy” inhibits action by policy makers. Concern for the environment is greater than the policy makers realize. The key to getting support for space solar power may be the growing awareness of the threat of rapid global environmental change. Scientists are extending their traditional role of theory and observation to emphasize the risks of global change. The risks provide the context for action by policy makers to move toward sustainable systems. The transition to power from space is responsive to the environmental concerns and the need to stabilize the Global environment and consequently the Earth's economic and social stability. Development of Space-Based Solar Power 29Renewable Energy28 2Cheap, reliable access to space is a key issue in making SBSP economically viable. The mass to be deployed will mandate a reusable launch system. Trades of the number of stages will be needed to optimize the efficiency. One evaluation may be the air launch concept being developed by Burt Rutan. It enables the launch of the upper stages above much of the sensible atmosphere. This reduces aerodynamic loads, but may be limited by a reasonable takeoff weight. Current assembly concepts have assumed construction in low earth orbit. After completion, the solar power satellite would be transferred to a higher orbit. Propulsion to accomplish this is a critical issue. One concept that has the specific impulse to make transfer practical is Variable Specific Impulse Magnetoplasma Rocket (VASIMR). NASA spinoff firm, the Ad Astra Rocket Company, has announced a key milestone in ground testing of its prototype plasma drive technology, The VASIMR "helicon first stage" - which generates the plasma for acceleration by the rest of the drive has achieved its full rated power of 30 kilowatts using Argon propellant, according to the company. This paves the way for further trials in which the ion-cyclotron second stage will get to boost the helicon plasma stream to the target power of 200 kW. The idea of the plasma drives is to use electric power to blast reaction mass (in this case Argon) from its rocket nozzles at a much higher speed than regular chemical rockets can achieve. This means that the carrying spacecraft gets a lot more acceleration from a given amount of fuel. A potential demonstration for VASIMR is maintaining the orbit of the international space station (ISS) without the need to burn large amounts of chemical rocket fuel. This serves as a demonstration of the transfer large structures between orbits. Since the solar power satellite was studied in the late 1970’s there have been many advancements in subsystem technology. These advances have included (a) improvements in photo-voltaic efficiency from about 10% (1970s) to more than 40% (2007); (b) increases in robotics capabilities from simple teleoperated manipulators in a few degrees of freedom (1970s) to fully autonomous robotics with insect-class intelligence and 30-100 degrees of freedom (2007); (c) increases in the efficiency of solid-state devices from around 20% (1970s) to as much as 70%-90% (2007); (d) improvements in materials for structures from simple aluminum (1970s) to advanced composites including nanotechnology composites (2007); (e) the application to large space structures; (f) high temperature super-conductors and many other technologies may be integrated into the design. Microwave beams are constant and conversion efficiencies high. They can be beamed at densities substantially lower than that of sunlight. This delivers more energy per area than terrestrial solar energy. The peak density of the beam can be significantly less than noon sunlight, and at the edge of the rectenna equivalent to the leakage of a microwave oven. This low energy density and choice of wavelength also means that biological effects are likely to be low. The safety of wild life wandering into the beam is not expected to be an issue. The physics of electromagnetic energy beaming is uncompromising. The size of the antenna makes microwave beaming unsuitable as a “secret” weapon. The distance from the 3geostationary belt is so great that beams diverge beyond the coherence and power concentration needed for a weapon. The beam is likely to be designed to require a pilot beam transmitted from the rectenna site. Absent the pilot signal, the system can be programmed to go into an incoherent mode. Concerns may also be addressed through an inspection regime. The likelihood of the beam wandering over a city is extremely low. Even if it occurred, it would not be a hazard. Wireless energy transfer by laser beam represents a different set of requirements. To achieve comparable efficiency, the beam must be more intense. The clouds in the atmosphere will reduce the transfer. The intense beam may produce a hazardous level to be avoided by aircraft and satellites. Still for the application to military power supply, it may be a manageable method. At present, the United States has very limited capabilities to build large structures, very large aperture antennas or very high power systems in orbit. The capability to control and maneuver these systems in space must be developed and demonstrated. Presently, the ability to translate large mass between Earth orbits will be required for deployment SBSP. Eventually, the capability for in-space manufacturing and construction or in-situ space resource utilization may be developed, but at this point it is a challenge that should not be incorporated into the program. One critical item to be demonstrated is capability for beamed power and application to propulsion of large space systems. The Thunderstorm Solar Power Satellite (TSPS) is a concept for interacting with thunderstorms to prevent formation of tornadoes. TSPS benefits are saving lives and reducing property. These benefits are not as sensitive to the system economics as the commercial solar power satellite and justify government investment in space solar power. The TSPS can develop and demonstrate the technology and operations critical to understanding the cost of space solar power. Consequently, there is no direct competition with fossil fuel based power supplies until SSP technology and operations have been demonstrated. Before weather modification can be safely attempted, the fine structure of thunderstorms must be simulated and related to tornadogenesis 3. Environmental Benefits Advocates of space solar power have been presenting the concepts as a means to help meet world energy needs. This argument has not been effective in garnering support for even basic research and technology development. Fossil fuel alternatives have been too cheap and near term effect on the “economy” inhibits action by policy makers. Concern for the environment is greater than the policy makers realize. The key to getting support for space solar power may be the growing awareness of the threat of rapid global environmental change. Scientists are extending their traditional role of theory and observation to emphasize the risks of global change. The risks provide the context for action by policy makers to move toward sustainable systems. The transition to power from space is responsive to the environmental concerns and the need to stabilize the Global environment and consequently the Earth's economic and social stability. Development of Space-Based Solar Power 29Renewable Energy30 4The "overview effect" from space has played a major role in developing a public sense of the fragile nature of the global environment. Stress on the Earth’s environmental system is increasing due to the buildup of carbon dioxide and other greenhouse gases. Models predicting the response to this buildup have not performed well in projecting the effect on the Earth’s climate because of the complexity of the system and the feed- backs within the system. Even the direction of climate change has not been predictable. An enhanced greenhouse effect has not been detected by temperature measurements. There may be interactions that are not well defined by the computer models, but that are reducing the stability of the Earth system. Because of the potential influence on the stability of the ocean currents that transport heat from the tropics to the higher latitudes, there is even a risk of returning the Earth to a glacial period rather than the global warming that is the present paradigm. Analysis of glacial ice cores indicates that such a shift can take place in less than a few decades. The likely effects of rapid climate change are increases in storm intensities, flooding, droughts, regional cropping shifts and sea level rise. These effects will have severe social and economic consequences. The rate of change and it’s direction leave civilization vulnerable to severe economic change in a period of significant population growth. Sustainable development has become the mantra for dealing with the potential global crises that are facing civilization. Clean, renewable energy is a resource that meets the criteria of sustainability. Collecting solar energy is prime candidate. Collecting the energy in space provides significant advantages in continuity of supply, although it’s development represents many challenges. A primary challenge is the issue of large initial cost prior to generating a return on that investment. The NASA Fresh Look at Space Solar Power study shows that concepts needing less initial investment are feasible. Even so, early SSP systems are not likely to be price competitive unless fossil fuel pricing incorporates the long range economic impact. The risks identified through the rigor of the U. S. Global Change Research Program (USGCRP) must provide the motivation for action toward sustainable systems. The USGCRP is an integrated program documenting the Earth system, understanding Earth system processes and developing computer models to predict the course of changes induced by humans or as the result of natural variations. The program is beginning to analyze the environmental, socioeconomic and health consequences of global change. The obvious next step is to assess means for mitigation of the effects of global change. The prosperity of future generations is dependent on a stable global environment. To ensure environmental stability, a continued effort to understand the effect of human activities must be a priority. Just understanding may not be sufficient because of the complex relationships of greenhouse gases, wind circulation, ocean currents and atmospheric water vapor. It is undisputed that carbon dioxide in the atmosphere has increased by over twenty percent since the beginning of the industrial age. Fossil fuels are certainly a major contributor to that increase. By replacing fossil fuel use, SSP could reduce the buildup of CO2 in the atmosphere and the consequent climate changes from an enhanced greenhouse effect. There 5are economic returns from a space-based power source that will lead to commercial management and operation of the system. There will continue to be an element of the political community that is committed to the short-term view because of the immediate economic impact. This reality is a factor that will have to be dealt with through facts and risk assessment for the long term view. The anticipated benefit to the Earth's environment is the overarching objective that may provide support for technology development and demonstration toward space solar power for use on Earth. 4. Summary Space-Based Solar Power is a huge project. It might be considered comparable in scale to the national railroads, highway system, or electrification project rather than the Manhattan or Apollo endeavors. However, unlike such purely national projects, this project also has components that are analogous to the development of the high�volume international civil aviation system. Such a large endeavor includes significant international and environmental implications. As such it would require a corresponding amount of political will to realize its benefits. Most of America’s spending in space does not provide any direct monetary revenue. SBSP will create new markets and produce new products. Great powers have historically succeeded by finding or inventing products and services not just to sell to themselves, but to sell to others. Today, investments in space are measured in billions of dollars. The energy market is trillions of dollars and will generate substantial new wealth for our nation and our world. Investments to develop SBSP have significant economic spin-offs. They open up or enable other new industries such as space industrial processes, space tourism, enhanced telecommunications, and use of off-world resources. After the fundamental technological risks have been defined, shifting SBSP from a research and development project to a financial and production program is needed. Several major challenges will need to be overcome to make SBSP a reality, including the creation of low� cost space access and a supporting infrastructure system on Earth and in space. The opportunity to export energy as the first marketable commodity from space will motivate commercial sector solutions to these challenges. The delivered commodity can be used for base�load terrestrial electrical power, wide�area broadcast power, carbon-neutral synthetic fuels production, military tactical support or as an in-space satellite energy utility. Solving these space access and operations challenges for SBSP will in turn also open space for space tourism, manufacturing, lunar or asteroid resource utilization, and eventually expansion of human presence and permanent settlement within our solar system. Space-based geoengineering concepts for environmental countermeasures are a potential supplement to earth-based actions. By defining options and benefits, SBSP may alert decision-makers to the potential of space operations as more than a tool to monitor the course of global change. Within the envelope of environmental protection is the preventing Development of Space-Based Solar Power 31Renewable Energy30 4The "overview effect" from space has played a major role in developing a public sense of the fragile nature of the global environment. Stress on the Earth’s environmental system is increasing due to the buildup of carbon dioxide and other greenhouse gases. Models predicting the response to this buildup have not performed well in projecting the effect on the Earth’s climate because of the complexity of the system and the feed- backs within the system. Even the direction of climate change has not been predictable. An enhanced greenhouse effect has not been detected by temperature measurements. There may be interactions that are not well defined by the computer models, but that are reducing the stability of the Earth system. Because of the potential influence on the stability of the ocean currents that transport heat from the tropics to the higher latitudes, there is even a risk of returning the Earth to a glacial period rather than the global warming that is the present paradigm. Analysis of glacial ice cores indicates that such a shift can take place in less than a few decades. The likely effects of rapid climate change are increases in storm intensities, flooding, droughts, regional cropping shifts and sea level rise. These effects will have severe social and economic consequences. The rate of change and it’s direction leave civilization vulnerable to severe economic change in a period of significant population growth. Sustainable development has become the mantra for dealing with the potential global crises that are facing civilization. Clean, renewable energy is a resource that meets the criteria of sustainability. Collecting solar energy is prime candidate. Collecting the energy in space provides significant advantages in continuity of supply, although it’s development represents many challenges. A primary challenge is the issue of large initial cost prior to generating a return on that investment. The NASA Fresh Look at Space Solar Power study shows that concepts needing less initial investment are feasible. Even so, early SSP systems are not likely to be price competitive unless fossil fuel pricing incorporates the long range economic impact. The risks identified through the rigor of the U. S. Global Change Research Program (USGCRP) must provide the motivation for action toward sustainable systems. The USGCRP is an integrated program documenting the Earth system, understanding Earth system processes and developing computer models to predict the course of changes induced by humans or as the result of natural variations. The program is beginning to analyze the environmental, socioeconomic and health consequences of global change. The obvious next step is to assess means for mitigation of the effects of global change. The prosperity of future generations is dependent on a stable global environment. To ensure environmental stability, a continued effort to understand the effect of human activities must be a priority. Just understanding may not be sufficient because of the complex relationships of greenhouse gases, wind circulation, ocean currents and atmospheric water vapor. It is undisputed that carbon dioxide in the atmosphere has increased by over twenty percent since the beginning of the industrial age. Fossil fuels are certainly a major contributor to that increase. By replacing fossil fuel use, SSP could reduce the buildup of CO2 in the atmosphere and the consequent climate changes from an enhanced greenhouse effect. There 5are economic returns from a space-based power source that will lead to commercial management and operation of the system. There will continue to be an element of the political community that is committed to the short-term view because of the immediate economic impact. This reality is a factor that will have to be dealt with through facts and risk assessment for the long term view. The anticipated benefit to the Earth's environment is the overarching objective that may provide support for technology development and demonstration toward space solar power for use on Earth. 4. Summary Space-Based Solar Power is a huge project. It might be considered comparable in scale to the national railroads, highway system, or electrification project rather than the Manhattan or Apollo endeavors. However, unlike such purely national projects, this project also has components that are analogous to the development of the high�volume international civil aviation system. Such a large endeavor includes significant international and environmental implications. As such it would require a corresponding amount of political will to realize its benefits. Most of America’s spending in space does not provide any direct monetary revenue. SBSP will create new markets and produce new products. Great powers have historically succeeded by finding or inventing products and services not just to sell to themselves, but to sell to others. Today, investments in space are measured in billions of dollars. The energy market is trillions of dollars and will generate substantial new wealth for our nation and our world. Investments to develop SBSP have significant economic spin-offs. They open up or enable other new industries such as space industrial processes, space tourism, enhanced telecommunications, and use of off-world resources. After the fundamental technological risks have been defined, shifting SBSP from a research and development project to a financial and production program is needed. Several major challenges will need to be overcome to make SBSP a reality, including the creation of low� cost space access and a supporting infrastructure system on Earth and in space. The opportunity to export energy as the first marketable commodity from space will motivate commercial sector solutions to these challenges. The delivered commodity can be used for base�load terrestrial electrical power, wide�area broadcast power, carbon-neutral synthetic fuels production, military tactical support or as an in-space satellite energy utility. Solving these space access and operations challenges for SBSP will in turn also open space for space tourism, manufacturing, lunar or asteroid resource utilization, and eventually expansion of human presence and permanent settlement within our solar system. Space-based geoengineering concepts for environmental countermeasures are a potential supplement to earth-based actions. By defining options and benefits, SBSP may alert decision-makers to the potential of space operations as more than a tool to monitor the course of global change. Within the envelope of environmental protection is the preventing Development of Space-Based Solar Power 31Renewable Energy32 6tornadoes concept. It promises early benefits by saving lives and reducing property damage. The principal payoff is projected to be the demonstration of space solar power technology and operations. This can lead to investment by the commercial energy organizations when their technical and operational risk is reduced. Once the potential for clean renewable energy from space is demonstrated, the way will be opened for further exploration and development of space. 5. References Stephen L. Klineberg, ‘Trends in Stakeholder Opinions: Findings from the Texas Environmental Survey (1990-1996)’, DeLange-Woodlands Conference, 1997. ‘Our Changing Planet, The FY 1997 U. S. Global Change Research Program’, Supplement to the President’s Fiscal Year 1997 Budget. National Security Space Office, “Space-Based Solar Power as an Opportunity for Strategic Security, Phase 0 Architecture Feasibility Study”, Report to Director, 10 October 2007. Dr. Bernard J. Eastlund, Lyle M. Jenkins, “Taming Tornadoes: Storm Abatement from Space”, IEEE Aerospace Conference, ID 205, March 7, 2001. "Policy Implications of Greenhouse Warming: Mitigation, Adaptation and the Science Base", National Academy Press, Washington, D. C., 1992. B.J. Eastlund, "Systems Considerations of Weather Modification Experiments Using High Power Electromagnetic Radiation", Workshop on Space Exploration and Resources Exploitation - Explospace, European Space Agency, October 20-22, 1998. Glaser, P. E.. Feasibility Study of a Satellite Solar Power Station, NASA Contact. Rep. CR-2357, NTIS N74-17784. Nat. Tech. Inform. Serv., Springfield, Va., USA,1974. M. Xue and K. K. Droegemeier. “The Advanced Regional Prediction System (ARPS)- A Multi-Scale Nonhydrostatic Atmospheric Simulation and Prediction Tool”, Model Dynamics, Monthly Weather Review. Eastlund, Bernard J., and Jenkins, Lyle M., “Space-based Concepts for Taming Tornadoes”, 51st International Astronautical Congress, Rio de Janeiro, Brazil October 2, 2000 Ming Sue, “Tornadogenesis within a Simulated Supercell Storm”, 22nd Severe Local Storms Conference, 6 October 2004. S. Businger, S. Chiswell, M. Bevis, J. Duan, R. Anthes, C. Rocken, R. Ware, M. Exner, T. van Hove, and S. Solheim - The Promise of GPS Atmospheric Monitoring, 1996 Mankins, John, “A Fresh Look at the Concept of Space Solar Power”, SPS ’97, Energy and Space for Humanity, August 26, 1997 Dr. Bernard J. Eastlund, Lyle M. Jenkins, “Taming Tornadoes: Storm Abatement from Space”, IEEE Aerospace Conference, ID 205, March 7, 2001. 7 Fig. 1. Top left-NASA/DOE reference 5GW; top center-NASA Sun Tower 200MW; top right-Integrated Symmetrical Concentrator; center left-JAXA Free Flyer Model; center right-USEF Tethered SPS; bottom- ESA Sail Tower 400MW Development of Space-Based Solar Power 33Renewable Energy32 6tornadoes concept. It promises early benefits by saving lives and reducing property damage. The principal payoff is projected to be the demonstration of space solar power technology and operations. This can lead to investment by the commercial energy organizations when their technical and operational risk is reduced. Once the potential for clean renewable energy from space is demonstrated, the way will be opened for further exploration and development of space. 5. References Stephen L. Klineberg, ‘Trends in Stakeholder Opinions: Findings from the Texas Environmental Survey (1990-1996)’, DeLange-Woodlands Conference, 1997. ‘Our Changing Planet, The FY 1997 U. S. Global Change Research Program’, Supplement to the President’s Fiscal Year 1997 Budget. National Security Space Office, “Space-Based Solar Power as an Opportunity for Strategic Security, Phase 0 Architecture Feasibility Study”, Report to Director, 10 October 2007. Dr. Bernard J. Eastlund, Lyle M. Jenkins, “Taming Tornadoes: Storm Abatement from Space”, IEEE Aerospace Conference, ID 205, March 7, 2001. "Policy Implications of Greenhouse Warming: Mitigation, Adaptation and the Science Base", National Academy Press, Washington, D. C., 1992. B.J. Eastlund, "Systems Considerations of Weather Modification Experiments Using High Power Electromagnetic Radiation", Workshop on Space Exploration and Resources Exploitation - Explospace, European Space Agency, October 20-22, 1998. Glaser, P. E.. Feasibility Study of a Satellite Solar Power Station, NASA Contact. Rep. CR-2357, NTIS N74-17784. Nat. Tech. Inform. Serv., Springfield, Va., USA,1974. M. Xue and K. K. Droegemeier. “The Advanced Regional Prediction System (ARPS)- A Multi-Scale Nonhydrostatic Atmospheric Simulation and Prediction Tool”, Model Dynamics, Monthly Weather Review. Eastlund, Bernard J., and Jenkins, Lyle M., “Space-based Concepts for Taming Tornadoes”, 51st International Astronautical Congress, Rio de Janeiro, Brazil October 2, 2000 Ming Sue, “Tornadogenesis within a Simulated Supercell Storm”, 22nd Severe Local Storms Conference, 6 October 2004. S. Businger, S. Chiswell, M. Bevis, J. Duan, R. Anthes, C. Rocken, R. Ware, M. Exner, T. van Hove, and S. Solheim - The Promise of GPS Atmospheric Monitoring, 1996 Mankins, John, “A Fresh Look at the Concept of Space Solar Power”, SPS ’97, Energy and Space for Humanity, August 26, 1997 Dr. Bernard J. Eastlund, Lyle M. Jenkins, “Taming Tornadoes: Storm Abatement from Space”, IEEE Aerospace Conference, ID 205, March 7, 2001. 7 Fig. 1. Top left-NASA/DOE reference 5GW; top center-NASA Sun Tower 200MW; top right-Integrated Symmetrical Concentrator; center left-JAXA Free Flyer Model; center right-USEF Tethered SPS; bottom- ESA Sail Tower 400MW Development of Space-Based Solar Power 33Renewable Energy34 8 F Fig. 2. Thunderstorm Solar Power Satellite Concept for preventing tornadoes. 9 Fig. 3. Computer simulation of tornadogenesis. Fig. 4. SBSP concept. Development of Space-Based Solar Power 35Renewable Energy34 8 F Fig. 2. Thunderstorm Solar Power Satellite Concept for preventing tornadoes. 9 Fig. 3. Computer simulation of tornadogenesis. Fig. 4. SBSP concept. Development of Space-Based Solar Power 35Renewable Energy36 Increasing the energy yield of generation from new and renewable energy resources 37Increasing the energy yield of generation from new and renewable energy resourcesSamuel C. E. Jupe, Andrea Michiorri and Philip C. TaylorX Increasing the energy yield of generation from new and renewable energy resources Samuel C. E. Jupe, Andrea Michiorri and Philip C. Taylor Durham University UK 1. Introduction The impetus of governments, on an international scale, to move towards low-carbon economy targets has brought about the proliferation of electricity (and heat) generation from new and renewable energy (RE) resources. This, coupled with increasing consumer energy demands, has caused distribution network operators (DNOs) to seek methods of increasing the utilisation of their existing power system assets. The increased utilisation of assets must be realised cautiously such that the security of supply to customers is not reduced, particularly when the age of distribution network assets is taken into account. A developer that is seeking to connect generation of significant capacity may be offered a firm connection by the DNO on the condition that an investment is made (by the developer) in the necessary network reinforcements. However, the developer may not be able to justify the expense of the required reinforcement and may negotiate a non-firm or ‘constrained’ connection agreement, whereby the generation installation is tripped off or constrained back under certain network operating conditions. Furthermore, difficulties may be encountered when attempting to gain permission to build network infrastructure, in order to accommodate new generation installations, due to planning problems and environmental objections (Fox-Penner, 2001). One potential solution or means of deferring these problems is the adoption of real-time thermal rating systems which have the potential, in certain circumstances, to be both less invasive and more cost effective when compared to network reinforcement options. Non-firm generation connections are expected to occur more frequently as network power flow congestion occurs. Therefore the deployment of a power output control system, informed by real-time thermal ratings, may be required to increase the energy yield of generation from new and RE resources. The stages in the development of an output control system for generation installations are illustrated. Section 2 provides a comprehensive literature review in order to provide the context for the research presented. Section 3 describes the assessment of the location of power flow congestion within the power system (due to the proliferation of generation from new and RE resources) so as to facilitate the targeted development of thermal models for thermally vulnerable components. This is achieved through the calculation of thermal vulnerability factors that relate power flow sensitivity factors (derived from governing 4Renewable Energy36 Increasing the energy yield of generation from new and renewable energy resources 37Increasing the energy yield of generation from new and renewable energy resourcesSamuel C. E. Jupe, Andrea Michiorri and Philip C. TaylorX Increasing the energy yield of generation from new and renewable energy resources Samuel C. E. Jupe, Andrea Michiorri and Philip C. Taylor Durham University UK 1. Introduction The impetus of governments, on an international scale, to move towards low-carbon economy targets has brought about the proliferation of electricity (and heat) generation from new and renewable energy (RE) resources. This, coupled with increasing consumer energy demands, has caused distribution network operators (DNOs) to seek methods of increasing the utilisation of their existing power system assets. The increased utilisation of assets must be realised cautiously such that the security of supply to customers is not reduced, particularly when the age of distribution network assets is taken into account. A developer that is seeking to connect generation of significant capacity may be offered a firm connection by the DNO on the condition that an investment is made (by the developer) in the necessary network reinforcements. However, the developer may not be able to justify the expense of the required reinforcement and may negotiate a non-firm or ‘constrained’ connection agreement, whereby the generation installation is tripped off or constrained back under certain network operating conditions. Furthermore, difficulties may be encountered when attempting to gain permission to build network infrastructure, in order to accommodate new generation installations, due to planning problems and environmental objections (Fox-Penner, 2001). One potential solution or means of deferring these problems is the adoption of real-time thermal rating systems which have the potential, in certain circumstances, to be both less invasive and more cost effective when compared to network reinforcement options. Non-firm generation connections are expected to occur more frequently as network power flow congestion occurs. Therefore the deployment of a power output control system, informed by real-time thermal ratings, may be required to increase the energy yield of generation from new and RE resources. The stages in the development of an output control system for generation installations are illustrated. Section 2 provides a comprehensive literature review in order to provide the context for the research presented. Section 3 describes the assessment of the location of power flow congestion within the power system (due to the proliferation of generation from new and RE resources) so as to facilitate the targeted development of thermal models for thermally vulnerable components. This is achieved through the calculation of thermal vulnerability factors that relate power flow sensitivity factors (derived from governing 4Renewable Energy38 alternating current (AC) power flow equations) to component steady-state thermal limits. Section 4 describes models for the steady-state assessment of power system component real-time thermal ratings. Industrial standard lumped parameter models are described for overhead lines, electric cables and power transformers. In a consistent manner, these models allow the influence of environmental conditions (such as wind speed) on component real-time thermal ratings to be assessed. Section 5 describes thermal state estimation techniques that allow the rating of components, which are not directly monitored within the power system, to be assessed. Thermal state estimations facilitate the precise and reliable assessment of environmental conditions whereby limited meteorological monitoring allows the thermal state of components within a wide area to be assessed. This may then be validated through the carefully selected monitoring of component operating temperatures. In Section 6, the power flow sensitivity factors are brought together with component real-time thermal ratings and candidate strategies are presented for the power output control of single or multiple generation installations. In Section 7, a case study is used to illustrate the developmental stages described above. In Section 8 the strengths and weaknesses of the proposed output control system for generation installations are discussed. The research described in this chapter forms part of a UK government part-funded project (Neumann et al, 2008) which aims to develop and deploy an online power output controller for wind generation installations through the exploitation of component real-time thermal properties. This is based on the concept that high power flows resulting from wind generation at high wind speeds could be accommodated since the same wind speed has a positive effect on component cooling mechanisms. In this project the control system compares component real-time thermal ratings with network power flows and produces set points that are fed back to the generation scheme operator for implementation, as shown in Figure 1. Fig. 1. Generation power output controller informed by real-time thermal ratings 2. Background Installed capacity assessments for generation from new and RE resources are the current research focus of numerous institutions in order to determine the impact of voltage regulations, operational economics, fault levels, losses and thermal limits as constraining parameters. Voltage limitations and the wind farm installed capacity relative to the system fault level in 33kV networks are considered (Dinic et al., 2006) and it is concluded that capacitive compensation can allow capacity maximisation within operational limits. The economics of generation connections to exploit multiple new and RE resources are considered (Currie et al., 2006) with a methodology that facilitates greater generation installed capacities. In order to manage power flows within prescribed voltage and thermal limits, operating margins are utilised with an active power output control technique termed ‘trim then trip’. An optimal power flow (OPF) technique is developed (Vovos et al., 2005) along with an iterative procedure to calculate the possible installed capacity of generation at nodes based on fault level limitations. The impact of increased generation installed capacities on electrical losses within the IEEE 34-node test network is examined (Mendez Quezada et al., 2006) and it is concluded that losses follow a U-shaped trajectory when plotted as a function of the generation penetration. An OPF formulation is presented (Harrison & Wallace, 2005) to determine the maximum generation installed capacity based on thermal limits and statutory voltage regulation. The ‘reverse load-ability’ methodology coupled with OPF software models generators as loads with a fixed power factor and creates an analysis tool that could allow additional constraints (such as fault-level limitations) to be incorporated into the formulation if necessary. Significant research has been carried out at the transmission level for real-time thermal rating applications. Research tends to focus on overhead lines which, due to their exposure to the environment, exhibit the greatest rating variability. A description of the cost and suitability of different uprating techniques for overhead lines is described (Stephen, 2004) taking into account different operating conditions. This work shows how real-time thermal ratings can be a more appropriate solution than network reinforcement when connecting new customers to the network who are able to curtail their generation output or reduce their power demand requirement at short notice. Similarly, experience regarding thermal uprating in the UK is reported (Hoffmann & Clark, 2004) where it was suggested that real-time thermal ratings could give overhead lines an average uprating of 5% for 50% of the year. An example of a real-time thermal rating application for transmission overhead lines of Red Eléctrica de España is described (Soto et al., 1998) where a limited number of weather stations are used to gather real-time data. The data is then processed using a meteorological model based on the Wind Atlas Analysis and Application Program (WAsP), taking into account the effect of obstacles and ground roughness, and the thermal rating is calculated. A similar system was developed in the USA by EPRI (Douglass & Edris, 1996) which considered overhead lines, power transformers, electric cables and substation equipment. Preliminary results of field tests (Douglass et al., 1997) show that up to 12 hours of low wind speeds (<0.76 ms-1) were observed during the field tests which therefore suggests that overhead line real-time thermal ratings may be lower than seasonal ratings for extended periods of time. Furthermore, a strong correlation was found to exist between independent air temperature measurements distributed along the lengths of the overhead lines. At the distribution level, a real-time thermal rating project carried out by the Dutch companies Increasing the energy yield of generation from new and renewable energy resources 39Renewable Energy38 alternating current (AC) power flow equations) to component steady-state thermal limits. Section 4 describes models for the steady-state assessment of power system component real-time thermal ratings. Industrial standard lumped parameter models are described for overhead lines, electric cables and power transformers. In a consistent manner, these models allow the influence of environmental conditions (such as wind speed) on component real-time thermal ratings to be assessed. Section 5 describes thermal state estimation techniques that allow the rating of components, which are not directly monitored within the power system, to be assessed. Thermal state estimations facilitate the precise and reliable assessment of environmental conditions whereby limited meteorological monitoring allows the thermal state of components within a wide area to be assessed. This may then be validated through the carefully selected monitoring of component operating temperatures. In Section 6, the power flow sensitivity factors are brought together with component real-time thermal ratings and candidate strategies are presented for the power output control of single or multiple generation installations. In Section 7, a case study is used to illustrate the developmental stages described above. In Section 8 the strengths and weaknesses of the proposed output control system for generation installations are discussed. The research described in this chapter forms part of a UK government part-funded project (Neumann et al, 2008) which aims to develop and deploy an online power output controller for wind generation installations through the exploitation of component real-time thermal properties. This is based on the concept that high power flows resulting from wind generation at high wind speeds could be accommodated since the same wind speed has a positive effect on component cooling mechanisms. In this project the control system compares component real-time thermal ratings with network power flows and produces set points that are fed back to the generation scheme operator for implementation, as shown in Figure 1. Fig. 1. Generation power output controller informed by real-time thermal ratings 2. Background Installed capacity assessments for generation from new and RE resources are the current research focus of numerous institutions in order to determine the impact of voltage regulations, operational economics, fault levels, losses and thermal limits as constraining parameters. Voltage limitations and the wind farm installed capacity relative to the system fault level in 33kV networks are considered (Dinic et al., 2006) and it is concluded that capacitive compensation can allow capacity maximisation within operational limits. The economics of generation connections to exploit multiple new and RE resources are considered (Currie et al., 2006) with a methodology that facilitates greater generation installed capacities. In order to manage power flows within prescribed voltage and thermal limits, operating margins are utilised with an active power output control technique termed ‘trim then trip’. An optimal power flow (OPF) technique is developed (Vovos et al., 2005) along with an iterative procedure to calculate the possible installed capacity of generation at nodes based on fault level limitations. The impact of increased generation installed capacities on electrical losses within the IEEE 34-node test network is examined (Mendez Quezada et al., 2006) and it is concluded that losses follow a U-shaped trajectory when plotted as a function of the generation penetration. An OPF formulation is presented (Harrison & Wallace, 2005) to determine the maximum generation installed capacity based on thermal limits and statutory voltage regulation. The ‘reverse load-ability’ methodology coupled with OPF software models generators as loads with a fixed power factor and creates an analysis tool that could allow additional constraints (such as fault-level limitations) to be incorporated into the formulation if necessary. Significant research has been carried out at the transmission level for real-time thermal rating applications. Research tends to focus on overhead lines which, due to their exposure to the environment, exhibit the greatest rating variability. A description of the cost and suitability of different uprating techniques for overhead lines is described (Stephen, 2004) taking into account different operating conditions. This work shows how real-time thermal ratings can be a more appropriate solution than network reinforcement when connecting new customers to the network who are able to curtail their generation output or reduce their power demand requirement at short notice. Similarly, experience regarding thermal uprating in the UK is reported (Hoffmann & Clark, 2004) where it was suggested that real-time thermal ratings could give overhead lines an average uprating of 5% for 50% of the year. An example of a real-time thermal rating application for transmission overhead lines of Red Eléctrica de España is described (Soto et al., 1998) where a limited number of weather stations are used to gather real-time data. The data is then processed using a meteorological model based on the Wind Atlas Analysis and Application Program (WAsP), taking into account the effect of obstacles and ground roughness, and the thermal rating is calculated. A similar system was developed in the USA by EPRI (Douglass & Edris, 1996) which considered overhead lines, power transformers, electric cables and substation equipment. Preliminary results of field tests (Douglass et al., 1997) show that up to 12 hours of low wind speeds (<0.76 ms-1) were observed during the field tests which therefore suggests that overhead line real-time thermal ratings may be lower than seasonal ratings for extended periods of time. Furthermore, a strong correlation was found to exist between independent air temperature measurements distributed along the lengths of the overhead lines. At the distribution level, a real-time thermal rating project carried out by the Dutch companies Increasing the energy yield of generation from new and renewable energy resources 39Renewable Energy40 NUON and KEMA (Nuijten & Geschiere, 2005) demonstrates the operating temperature monitoring of overhead lines, electric cables and power transformers. The advantages of a real-time thermal rating system for the connection of generation from new and RE resources are reported in various sources, each of which considers only single power system components. It is demonstrated (Helmer, 2000) that the rating of transformers positioned at the base of wind turbines may presently be oversized by up to 20%. Moreover, the power flowing in an overhead line close to a wind farm is compared to its real-time thermal rating using WAsP (Belben & Ziesler, 2002). In this research it was highlighted that high power flows resulting from wind generation at high wind speeds could be accommodated since the same wind speed has a positive effect on the line cooling. This observation makes the adoption of real-time thermal rating systems relevant in applications where strong correlations exist between the cooling effect of environmental conditions and electrical power flow transfers. Moreover, the influence of component thermal model input errors on the accuracy of real-time thermal rating systems is studied (Piccolo et al., 2004; Ippolito et al., 2004; Villacci & Vaccaro, 2007). The application of different state estimation techniques, such as affine arithmetic, interval arithmetic and Montecarlo simulations was studied for overhead lines, electric cables and power transformers. Errors of up to ±20% for an operating point of 75oC, ±29% for an operating point of 60oC and ±15% for an operating point of 65oC were found when estimating the operating temperature of overhead lines, electric cables and power transformers respectively. This highlights the necessity to have reliable and accurate environmental condition monitoring. The thermal models, used to estimate real-time thermal ratings for different types of power system components, are fundamental to this research as the accuracy of the models influence significantly the accuracy of real-time thermal ratings obtained. Particular attention was given to industrial standards because of their wide application and validation both in industry and academia. For overhead lines, the models (House & Tuttle, 1959; Morgan, 1982) have been developed into industrial standards by the IEC (IEC, 1995), CIGRE (WG 22.12, 1992) and IEEE (IEEE, 1993). Static seasonal ratings for different standard conductors and for calculated risks are provided by the Electricity Network Association (ENA, 1986). Thermal model calculation methods for electric cable ratings are described (Neher & McGrath, 1957) and developed into an industrial standard by the IEC (IEC, 1994). The same models are used by the IEEE (IEEE, 1994) and the ENA (ENA, 2004) to produce tables of calculated ratings for particular operating conditions. Power transformer thermal behaviour is described (Susa et al., 2005) with further models described in the industrial standards by the IEC (IEC, 2008), the IEEE (ANSI/IEEE, 1981) and the ENA (ENA, 1971). The work detailed in this chapter moves beyond the offline assessment of generation installed capacities to outline the development stages in the online power output control of generation installations. The thermal vulnerability factor assessments presented in this chapter complement network characterisation methods (Berende et al., 2005) by first identifying the type (overhead line, electric cable, power transformer) and geographical location of thermally vulnerable components. The assessments may be used to give a holistic network view of the impact of multiple generation installations in concurrent operation on accumulated power flows and hence vulnerable component locations. This facilitates the targeted development of component thermal models. Moreover, (Michiorri et al., 2009) describes the influence of environmental conditions on multiple power system component types simultaneously. This is of particular relevance in situations where the increased power flow resulting from the alleviation of the thermal constraint on one power system component may cause an entirely different component to constrain power flows. Whilst OPF is acknowledged as a powerful tool for the offline planning of electrical networks, there is an emerging requirement to manage non-firm generation connections in an online manner. This requires the deployment of a system which has the capability of utilising real-time information about the thermal status of the network and, in reaching a control decision, guarantees that the secure operation of the distribution network is maintained. The rapid processing time, reduced memory requirements and robustness associated with embedding predetermined power flow sensitivity factors in a power output control system for generation installations make it attractive for substation and online applications. This is strengthened further by the ability of the power output control system to readily integrate component real-time thermal ratings in the management of network power flows for increased new and renewable energy yields. Moreover, since this research project aims to develop and deploy an economically viable real-time thermal rating system, it is important that algorithms are developed with fast computational speeds using limited environmental condition monitoring. Thus an inverse distance interpolation technique is used for modelling environmental conditions across a wide geographical area, which offers faster computational speeds than applications such as WAsP. Beyond the research described above, this chapter also suggests potential annual energy yields that may be gained through the deployment of an output control system for generation installations. 3. Power flow sensitivity factors Once the inverse Jacobian has been evaluated in the full AC power flow solution, perturbations about a given set of system conditions may be calculated using Eq.1 (Wood & Wollenberg, 1996). This gives the changes expected in bus voltage angles and voltage magnitudes due to injections of real or reactive power.   ΔQΔPΔQΔP JVΔΔθVΔΔθkkii1-kkii (1) Where θi and θk (rad) represent voltage angles at nodes i and k respectively, |Vi| and |Vk| (kV) represent nodal voltages, J is the Jacobian matrix, Pi and Pk (MW) represent real power injections at nodes i and k respectively and Qi and Qk (MVAr) represent reactive power injections at nodes i and k respectively. The work presented in this paper is specifically concerned with calculating the effect of a perturbation of ΔPm – that is an injection of power at unity power factor (real power) into node m. Since the generation shifts, the reference (slack) bus compensates for the increase in Increasing the energy yield of generation from new and renewable energy resources 41Renewable Energy40 NUON and KEMA (Nuijten & Geschiere, 2005) demonstrates the operating temperature monitoring of overhead lines, electric cables and power transformers. The advantages of a real-time thermal rating system for the connection of generation from new and RE resources are reported in various sources, each of which considers only single power system components. It is demonstrated (Helmer, 2000) that the rating of transformers positioned at the base of wind turbines may presently be oversized by up to 20%. Moreover, the power flowing in an overhead line close to a wind farm is compared to its real-time thermal rating using WAsP (Belben & Ziesler, 2002). In this research it was highlighted that high power flows resulting from wind generation at high wind speeds could be accommodated since the same wind speed has a positive effect on the line cooling. This observation makes the adoption of real-time thermal rating systems relevant in applications where strong correlations exist between the cooling effect of environmental conditions and electrical power flow transfers. Moreover, the influence of component thermal model input errors on the accuracy of real-time thermal rating systems is studied (Piccolo et al., 2004; Ippolito et al., 2004; Villacci & Vaccaro, 2007). The application of different state estimation techniques, such as affine arithmetic, interval arithmetic and Montecarlo simulations was studied for overhead lines, electric cables and power transformers. Errors of up to ±20% for an operating point of 75oC, ±29% for an operating point of 60oC and ±15% for an operating point of 65oC were found when estimating the operating temperature of overhead lines, electric cables and power transformers respectively. This highlights the necessity to have reliable and accurate environmental condition monitoring. The thermal models, used to estimate real-time thermal ratings for different types of power system components, are fundamental to this research as the accuracy of the models influence significantly the accuracy of real-time thermal ratings obtained. Particular attention was given to industrial standards because of their wide application and validation both in industry and academia. For overhead lines, the models (House & Tuttle, 1959; Morgan, 1982) have been developed into industrial standards by the IEC (IEC, 1995), CIGRE (WG 22.12, 1992) and IEEE (IEEE, 1993). Static seasonal ratings for different standard conductors and for calculated risks are provided by the Electricity Network Association (ENA, 1986). Thermal model calculation methods for electric cable ratings are described (Neher & McGrath, 1957) and developed into an industrial standard by the IEC (IEC, 1994). The same models are used by the IEEE (IEEE, 1994) and the ENA (ENA, 2004) to produce tables of calculated ratings for particular operating conditions. Power transformer thermal behaviour is described (Susa et al., 2005) with further models described in the industrial standards by the IEC (IEC, 2008), the IEEE (ANSI/IEEE, 1981) and the ENA (ENA, 1971). The work detailed in this chapter moves beyond the offline assessment of generation installed capacities to outline the development stages in the online power output control of generation installations. The thermal vulnerability factor assessments presented in this chapter complement network characterisation methods (Berende et al., 2005) by first identifying the type (overhead line, electric cable, power transformer) and geographical location of thermally vulnerable components. The assessments may be used to give a holistic network view of the impact of multiple generation installations in concurrent operation on accumulated power flows and hence vulnerable component locations. This facilitates the targeted development of component thermal models. Moreover, (Michiorri et al., 2009) describes the influence of environmental conditions on multiple power system component types simultaneously. This is of particular relevance in situations where the increased power flow resulting from the alleviation of the thermal constraint on one power system component may cause an entirely different component to constrain power flows. Whilst OPF is acknowledged as a powerful tool for the offline planning of electrical networks, there is an emerging requirement to manage non-firm generation connections in an online manner. This requires the deployment of a system which has the capability of utilising real-time information about the thermal status of the network and, in reaching a control decision, guarantees that the secure operation of the distribution network is maintained. The rapid processing time, reduced memory requirements and robustness associated with embedding predetermined power flow sensitivity factors in a power output control system for generation installations make it attractive for substation and online applications. This is strengthened further by the ability of the power output control system to readily integrate component real-time thermal ratings in the management of network power flows for increased new and renewable energy yields. Moreover, since this research project aims to develop and deploy an economically viable real-time thermal rating system, it is important that algorithms are developed with fast computational speeds using limited environmental condition monitoring. Thus an inverse distance interpolation technique is used for modelling environmental conditions across a wide geographical area, which offers faster computational speeds than applications such as WAsP. Beyond the research described above, this chapter also suggests potential annual energy yields that may be gained through the deployment of an output control system for generation installations. 3. Power flow sensitivity factors Once the inverse Jacobian has been evaluated in the full AC power flow solution, perturbations about a given set of system conditions may be calculated using Eq.1 (Wood & Wollenberg, 1996). This gives the changes expected in bus voltage angles and voltage magnitudes due to injections of real or reactive power.   ΔQΔPΔQΔP JVΔΔθVΔΔθkkii1-kkii (1) Where θi and θk (rad) represent voltage angles at nodes i and k respectively, |Vi| and |Vk| (kV) represent nodal voltages, J is the Jacobian matrix, Pi and Pk (MW) represent real power injections at nodes i and k respectively and Qi and Qk (MVAr) represent reactive power injections at nodes i and k respectively. The work presented in this paper is specifically concerned with calculating the effect of a perturbation of ΔPm – that is an injection of power at unity power factor (real power) into node m. Since the generation shifts, the reference (slack) bus compensates for the increase in Increasing the energy yield of generation from new and renewable energy resources 41Renewable Energy42 power. The Δθ and Δ|V| values in Eq.2 are thus equal to the derivative of the bus angles and voltage magnitudes with respect to a change in power at bus m.   ΔQΔPΔQΔP JVΔΔθrefrefmm1- (2) Thus the sensitivity factors for a real power injection at node m are given in Eq.3-6:    m,Pim,Pkk,im,Pk,idGddGdPdGdP:f (3) m,Pim,Pkk,im,Pk,idGVddGVdVPdGdP:)V(f (4)    m,Pim,Pkk,im,Pk,idGddGdQdGdQ:f (5)   m,Pim,Pkk,im,Pk,idGVddGVdVQdGdQ:Vf (6) Where f(θ) and f(V) represent functions of voltage angles and voltage magnitudes respectively, (∂P/∂θ)i,k, (∂P/∂V)i,k, (∂Q/∂θ)i,k and (∂Q/∂V)i,k, and represent elements within the Jacobian matrix and dθk/dGP,m, dθi/dGP,m, dVk/dGP,m and dVi/dGP,m represent elements corresponding to the relevant Δθ and Δ|V| values evaluated in Eq.2. This gives an overall power flow sensitivity factor (Si,k,m) in the component from node i to node k, due to an injection of real power, at node m, as in Eq.7.   m,Piim,Pkkk,im,Pim,Pkk,im,Piim,Pkkk,im,Pim,Pkk,icm,k,idGEEddGEEdEEQdGddGdQ j dGEEddGEEdEEPdGddGdPSSF (7) Simplified versions of the power flow sensitivity factor theory (focusing on the P-θ sensitivity) are used at the transmission level for real power flow sensitivity analyses. The generation shift factor (GSF) technique (Wood & Wollenberg, 1996) is acceptable for use in DC representations of AC systems where the network behaviour is approximated by neglecting MVAr flow and assuming voltage to be constant. However, in distribution networks those assumptions do not always hold since, in some cases, the electrical reactance (RE) and electrical resistance (X) of components is approximately equal (i.e. X/RE ≈ 1). Thus reactive power flow may contribute to a significant portion of the resultant power flowing in components. In these situations it is important that both real and reactive power flows are considered when assessing the locations of thermally vulnerable components and developing techniques for the online power output control of generation from new and RE resources. 3.1 Thermal vulnerability factors Eq.7 may be combined with the relevant component thermal rating and the resulting thermal vulnerability factor, as given in Eq.8, is standardised by conversion to a per unit term on the base MVA limm,k,im,k,i SSSFTVF on Sbase (8) where TVFi,k,m represents the thermal vulnerability factor of the component from node i to node k due to a real power injection at node m, SSFi,k,m represents the power flow sensitivity factor in the component from node i to node k, due to a real power injection at node m, Slim (MVA) represents the thermal limit of the component and Sbase is a predefined MVA base. This gives a consistent measure of component thermal vulnerabilities, relative to one another and accounts for different nodal real power injections, for a particular network operating condition. It can also be seen in Eq.9 that the sensitivity factor relative to the component rating is equivalent to the change in utilisation of a particular component from node i to node k, due to an injection of real power at node m P, mi,klimP, mi,klimi,k,mΔGΔUSΔGΔSSSSF (9) where SSFi,k,m represents the power flow sensitivity factor of the component, from node i to node k, due to a real power injection at node m, Slim (MVA) represents the thermal limit of component, ΔSi,k (MVA) represents the change in apparent power flow in the component from node i to node k, ΔGP,m (MW) represents the change in real power injection at node m and ΔUi,k represents the change in capacity utilisation of the component from node i to node k. Power flow sensitivity factors indicate the extent to which power flow changes within components due to nodal power injections. However, a large change in power flow, indicated by high sensitivity, does not necessarily mean a component is thermally vulnerable unless its rating is taken into account. A large power flow change in a component with a large thermal rating could be less critical than a small power flow change in a component with a small rating. By calculating the apparent power sensitivity relative to Increasing the energy yield of generation from new and renewable energy resources 43Renewable Energy42 power. The Δθ and Δ|V| values in Eq.2 are thus equal to the derivative of the bus angles and voltage magnitudes with respect to a change in power at bus m.   ΔQΔPΔQΔP JVΔΔθrefrefmm1- (2) Thus the sensitivity factors for a real power injection at node m are given in Eq.3-6:    m,Pim,Pkk,im,Pk,idGddGdPdGdP:f (3) m,Pim,Pkk,im,Pk,idGVddGVdVPdGdP:)V(f (4)    m,Pim,Pkk,im,Pk,idGddGdQdGdQ:f (5)   m,Pim,Pkk,im,Pk,idGVddGVdVQdGdQ:Vf (6) Where f(θ) and f(V) represent functions of voltage angles and voltage magnitudes respectively, (∂P/∂θ)i,k, (∂P/∂V)i,k, (∂Q/∂θ)i,k and (∂Q/∂V)i,k, and represent elements within the Jacobian matrix and dθk/dGP,m, dθi/dGP,m, dVk/dGP,m and dVi/dGP,m represent elements corresponding to the relevant Δθ and Δ|V| values evaluated in Eq.2. This gives an overall power flow sensitivity factor (Si,k,m) in the component from node i to node k, due to an injection of real power, at node m, as in Eq.7.   m,Piim,Pkkk,im,Pim,Pkk,im,Piim,Pkkk,im,Pim,Pkk,icm,k,idGEEddGEEdEEQdGddGdQ j dGEEddGEEdEEPdGddGdPSSF (7) Simplified versions of the power flow sensitivity factor theory (focusing on the P-θ sensitivity) are used at the transmission level for real power flow sensitivity analyses. The generation shift factor (GSF) technique (Wood & Wollenberg, 1996) is acceptable for use in DC representations of AC systems where the network behaviour is approximated by neglecting MVAr flow and assuming voltage to be constant. However, in distribution networks those assumptions do not always hold since, in some cases, the electrical reactance (RE) and electrical resistance (X) of components is approximately equal (i.e. X/RE ≈ 1). Thus reactive power flow may contribute to a significant portion of the resultant power flowing in components. In these situations it is important that both real and reactive power flows are considered when assessing the locations of thermally vulnerable components and developing techniques for the online power output control of generation from new and RE resources. 3.1 Thermal vulnerability factors Eq.7 may be combined with the relevant component thermal rating and the resulting thermal vulnerability factor, as given in Eq.8, is standardised by conversion to a per unit term on the base MVA limm,k,im,k,i SSSFTVF on Sbase (8) where TVFi,k,m represents the thermal vulnerability factor of the component from node i to node k due to a real power injection at node m, SSFi,k,m represents the power flow sensitivity factor in the component from node i to node k, due to a real power injection at node m, Slim (MVA) represents the thermal limit of the component and Sbase is a predefined MVA base. This gives a consistent measure of component thermal vulnerabilities, relative to one another and accounts for different nodal real power injections, for a particular network operating condition. It can also be seen in Eq.9 that the sensitivity factor relative to the component rating is equivalent to the change in utilisation of a particular component from node i to node k, due to an injection of real power at node m P, mi,klimP, mi,klimi,k,mΔGΔUSΔGΔSSSSF (9) where SSFi,k,m represents the power flow sensitivity factor of the component, from node i to node k, due to a real power injection at node m, Slim (MVA) represents the thermal limit of component, ΔSi,k (MVA) represents the change in apparent power flow in the component from node i to node k, ΔGP,m (MW) represents the change in real power injection at node m and ΔUi,k represents the change in capacity utilisation of the component from node i to node k. Power flow sensitivity factors indicate the extent to which power flow changes within components due to nodal power injections. However, a large change in power flow, indicated by high sensitivity, does not necessarily mean a component is thermally vulnerable unless its rating is taken into account. A large power flow change in a component with a large thermal rating could be less critical than a small power flow change in a component with a small rating. By calculating the apparent power sensitivity relative to Increasing the energy yield of generation from new and renewable energy resources 43Renewable Energy44 rating for each component, the thermally vulnerable components are identified and can be ranked for single nodal power injections or accumulated for multiple injections. 3.2 Factor assessments An empirical procedure (Jupe & Taylor, 2009b) to assess power flow sensitivity factors and generate lists of thermally vulnerable components for different network topologies has been developed as follows: Initially a ‘base case’ AC load flow is run in the power system simulation package to establish real, reactive and apparent power flows for each component. The procedure iterates by injecting 1pu of real power at each node of interest and recording the new component power flows. The initial flow, final flow and thermal rating of each component are used to relate component power flow sensitivity factors to nodal injections and ratings. The resulting power flow sensitivity factors and thermal vulnerability factors are efficiently stored in matrix form and, with the thermal vulnerability factors represented graphically, a visual identification of the most thermally vulnerable components is given. 4. Thermal modelling approach In order to assess, in a consistent manner, component real-time thermal ratings due to the influence of environmental conditions, thermal models were developed based on IEC standards for overhead lines (IEC, 1995), electric cables (IEC, 1994) and power transformers (IEC, 2008). Where necessary, refinements were made to the models (WG22.12, 1992; ENA, 2004). Steady-state models have been used in preference to dynamic models since this would provide a maximum allowable rating for long term power system operation. 4.1 Overhead lines Overhead line ratings are constrained by a necessity to maintain statutory clearances between the conductor and other objects. The temperature rise causes conductor elongation which, in turn, causes an increase in sag. The line sag, ψ (m), depends on the tension, H (N), the weight, mg (N) applied to the conductor inclusive of the dynamic force of the wind and the length of the span, L (m). The sag can be calculated as a catenary or its parabolic approximation, as given in Eq.10. H8mgL1H2mgLcoshmgH 2 (10) To calculate the tension, it is necessary to consider the thermal-tensional equilibrium of the conductor, as shown in Eq.11, where E represents the Young’s modulus of the conductor (Pa), A represents the cross-sectional area of the conductor (m2) and β represents the conductor’s thermal expansion coefficient (K-1).   222222212122211,c2,c HH24EALgmHH24EALgmTTEA  (11) For calculating the conductor operating temperature at a given current, or the maximum current for a given operating temperature, it is necessary to solve the energy balance between the heat dissipated in the conductor by the current, and the thermal exchange on its surface, as given in Eq.12 E2src RIqqq  (12) where qc represents convective heat exchange (Wm-1), qr represents radiative heat exchange (Wm-1), qs represents solar radiation (Wm-1) and I2RE represents the heat dissipated in the conductor due to the Joule effect (Wm-1). The proposed formulae (IEC, 1995) were used for the calculation of the contribution of solar radiation, radiative heat exchange and convective heat exchange as given in Eq.13-15 respectively, SrDqs  (13)   DTTq 4a4cSBr  (14)  acc TTNuq  (15) where α represents the absorption coefficient, D represents the external diameter of the conductor (m) and Sr represents solar radiation (Wm-2), ε represents the emission coefficient, σSB represents the Stefan-Boltzmann constant (Wm-2K-4), and Tc and Ta represent the respective conductor and ambient temperatures (K), Nu represents the Nusselt number and λ represents the air thermal conductivity (Wm-1K-1). The influences of wind direction and natural convection on convective heat exchange are not considered in the IEC standard model (IEC, 1995). However, in this research these effects were considered to be important, particularly as a wind direction perpendicular to the conductor would maximise the turbulence around the conductor and hence the heat exchange on its surface whereas a wind direction parallel to the conductor would reduce the heat exchange with respect to perpendicular wind direction. Therefore the modifications (WG22.12, 1992) given in Eq.16 and Eq.19 were used. It is possible to calculate the Nusselt number, Nu, from the Reynolds number, Re, as shown in Eq.17. The Reynolds number can be calculated using Eq.18   3,dirK2,dir1,dirdir WdsinKKK  (16)  61.02.0dir Re23.0Re65.0KNu  (17) 78.1ac92TTDWs10644.1Re  (18) where Kdir represents the wind direction influence coefficient, Kdir,1-3 represent wind direction coefficient constants, Wd represents the wind-conductor angle (rad) and Ws represents the wind speed (ms-1). Increasing the energy yield of generation from new and renewable energy resources 45Renewable Energy44 rating for each component, the thermally vulnerable components are identified and can be ranked for single nodal power injections or accumulated for multiple injections. 3.2 Factor assessments An empirical procedure (Jupe & Taylor, 2009b) to assess power flow sensitivity factors and generate lists of thermally vulnerable components for different network topologies has been developed as follows: Initially a ‘base case’ AC load flow is run in the power system simulation package to establish real, reactive and apparent power flows for each component. The procedure iterates by injecting 1pu of real power at each node of interest and recording the new component power flows. The initial flow, final flow and thermal rating of each component are used to relate component power flow sensitivity factors to nodal injections and ratings. The resulting power flow sensitivity factors and thermal vulnerability factors are efficiently stored in matrix form and, with the thermal vulnerability factors represented graphically, a visual identification of the most thermally vulnerable components is given. 4. Thermal modelling approach In order to assess, in a consistent manner, component real-time thermal ratings due to the influence of environmental conditions, thermal models were developed based on IEC standards for overhead lines (IEC, 1995), electric cables (IEC, 1994) and power transformers (IEC, 2008). Where necessary, refinements were made to the models (WG22.12, 1992; ENA, 2004). Steady-state models have been used in preference to dynamic models since this would provide a maximum allowable rating for long term power system operation. 4.1 Overhead lines Overhead line ratings are constrained by a necessity to maintain statutory clearances between the conductor and other objects. The temperature rise causes conductor elongation which, in turn, causes an increase in sag. The line sag, ψ (m), depends on the tension, H (N), the weight, mg (N) applied to the conductor inclusive of the dynamic force of the wind and the length of the span, L (m). The sag can be calculated as a catenary or its parabolic approximation, as given in Eq.10. H8mgL1H2mgLcoshmgH 2 (10) To calculate the tension, it is necessary to consider the thermal-tensional equilibrium of the conductor, as shown in Eq.11, where E represents the Young’s modulus of the conductor (Pa), A represents the cross-sectional area of the conductor (m2) and β represents the conductor’s thermal expansion coefficient (K-1).   222222212122211,c2,c HH24EALgmHH24EALgmTTEA  (11) For calculating the conductor operating temperature at a given current, or the maximum current for a given operating temperature, it is necessary to solve the energy balance between the heat dissipated in the conductor by the current, and the thermal exchange on its surface, as given in Eq.12 E2src RIqqq  (12) where qc represents convective heat exchange (Wm-1), qr represents radiative heat exchange (Wm-1), qs represents solar radiation (Wm-1) and I2RE represents the heat dissipated in the conductor due to the Joule effect (Wm-1). The proposed formulae (IEC, 1995) were used for the calculation of the contribution of solar radiation, radiative heat exchange and convective heat exchange as given in Eq.13-15 respectively, SrDqs  (13)   DTTq 4a4cSBr  (14)  acc TTNuq  (15) where α represents the absorption coefficient, D represents the external diameter of the conductor (m) and Sr represents solar radiation (Wm-2), ε represents the emission coefficient, σSB represents the Stefan-Boltzmann constant (Wm-2K-4), and Tc and Ta represent the respective conductor and ambient temperatures (K), Nu represents the Nusselt number and λ represents the air thermal conductivity (Wm-1K-1). The influences of wind direction and natural convection on convective heat exchange are not considered in the IEC standard model (IEC, 1995). However, in this research these effects were considered to be important, particularly as a wind direction perpendicular to the conductor would maximise the turbulence around the conductor and hence the heat exchange on its surface whereas a wind direction parallel to the conductor would reduce the heat exchange with respect to perpendicular wind direction. Therefore the modifications (WG22.12, 1992) given in Eq.16 and Eq.19 were used. It is possible to calculate the Nusselt number, Nu, from the Reynolds number, Re, as shown in Eq.17. The Reynolds number can be calculated using Eq.18   3,dirK2,dir1,dirdir WdsinKKK  (16)  61.02.0dir Re23.0Re65.0KNu  (17) 78.1ac92TTDWs10644.1Re  (18) where Kdir represents the wind direction influence coefficient, Kdir,1-3 represent wind direction coefficient constants, Wd represents the wind-conductor angle (rad) and Ws represents the wind speed (ms-1). Increasing the energy yield of generation from new and renewable energy resources 45Renewable Energy46 For null wind speeds, the Nusselt number must be calculated as in Eq.19 where Gr is the Grashof number, calculated as in Eq.20, and Pr is the Prandtl number   2,natK1,nat PrGrKNu  (19)    2acac32TTgTTDGr (20) where Knat,1-2 represent natural convection coefficients and υ represents kinematic viscosity (m2s-1). It should be noted that for wind speeds between 0-0.5ms-1 the larger of the Nusselt numbers resulting from Eq.17 and Eq.19 should be used. 4.2 Electric cables The current carrying capacity (ampacity) of electric cables is limited by the maximum operating temperature of the insulation. Sustained high currents may generate temperatures in exceedance of the maximum operating temperature, causing irreversible damage to the cable. In extreme cases this may result in complete insulation deterioration and cable destruction. The conductor temperature in steady-state conditions was modelled (Neher & McGrath, 1957; IEC, 1994; IEEE, 1994) to account for the heat balance between the power dissipated in the conductor by the Joule effect, and the heat dissipated in the environment through the thermal resistance, RT (mKW-1), of the insulation and the soil, due to the temperature difference, ΔT (K), as shown in Eq.21. The electrical current rating, I (A), may then be calculated, as shown in Eq.22. TE2 RTRI  (21) TE RRTI (22) Refinements incorporating dielectric losses, qd (Wm-1), the number of conductors in the cable, n, eddy currents and circulating currents in metallic sheaths, (λ1,2), electrical resistance variation with conductor temperature, RE(Tc) (Ω), skin and proximity effects and the thermal resistance of each insulating layer, RT,1-4, of lead to the more complex expression for calculating the cable ampacity as given in Eq.23.   )]RR)(1(nR)1(nR[TR)]RRR(nR21[qTI4,T3,T212,T11,TcE4,T3,T2,T1,Td (23) Thermal resistances, RT,1-3, for cylindrical layers are calculated using Eq.24 and soil thermal resistance, RT,4, is modelled using Eq.25    ddD21ln2R Ts3,2,1T (24)   1Dz2Dz2ln2R2bbTs4T (25) where ρs-T represents the soil thermal resistivity (Wm-1K-1), D and d represent the respective external and internal diameters (m) for the calculated layer and zb represents the burial depth of the cable (m). Other calculation methods (IEC, 1994) have to be utilised when operating conditions differ from those stated above (for example when the cable is in a duct or in open air). The model described above requires detailed knowledge of the electric cable installation. However, this information may not always be available and therefore it is difficult to make practical use of the model. In these circumstances an alternative model given in Eq.26 may be used (ENA, 2004). The rated current of electric cables, I0 (A), is given in tables depending on the nominal voltage level, V (kV), the standardised cable cross-sectional area, A (m2) and laying conditions (trefoil, flat formation; in air, in ducts or directly buried). The dependence of the cable ampacity on the actual soil temperature, Ts (K) away from the rated soil temperature Ts rated (K) as well as the actual soil thermal resistivity away from the rated soil thermal resistivity is made linear through the coefficients ξT and ξρ respectively.        rated T,sT,srated ssT0 TTlaying,V,AII   (26) Since this research concerns the influence of environmental conditions on component ratings, the effect of the nominal voltage level which influences the dielectric loss, qd in Eq.23 is not considered. The effect of the heating given by adjacent components is also neglected as it is assumed that each cable has already been de-rated to take this effect into account. 4.3 Power transformers The thermal model (IEC, 2008) can be used to calculate the winding hot spot temperature for power transformers. This is the most important parameter since hotspot temperature exceedance can damage the transformer in two ways. Firstly, a temperature exceedance of 120ºC-140ºC can induce the formation of bubbles in the coolant oil, which in turn is liable to cause an insulation breakdown due to the local reduction of dielectric insulation strength. Secondly, high temperatures increase the ageing rate of the winding insulation. For this reason the maximum operating temperature should not exceed the rated value. The thermal model consists of a heat balance between the power dissipated in the winding and iron core, and the heat transferred to the environment via the refrigerating circuit. Considering the ambient temperature, Ta (K), thermal resistance between the windings and the oil, RT,W (mKW-1), the thermal resistance between the heat exchanger and the air, RT,HE (mKW-1) and the power dissipated into the core, I2RE, windings (W), it is possible to calculate the hot spot temperature, THS (K) as in Eq.27. Increasing the energy yield of generation from new and renewable energy resources 47Renewable Energy46 For null wind speeds, the Nusselt number must be calculated as in Eq.19 where Gr is the Grashof number, calculated as in Eq.20, and Pr is the Prandtl number   2,natK1,nat PrGrKNu  (19)    2acac32TTgTTDGr (20) where Knat,1-2 represent natural convection coefficients and υ represents kinematic viscosity (m2s-1). It should be noted that for wind speeds between 0-0.5ms-1 the larger of the Nusselt numbers resulting from Eq.17 and Eq.19 should be used. 4.2 Electric cables The current carrying capacity (ampacity) of electric cables is limited by the maximum operating temperature of the insulation. Sustained high currents may generate temperatures in exceedance of the maximum operating temperature, causing irreversible damage to the cable. In extreme cases this may result in complete insulation deterioration and cable destruction. The conductor temperature in steady-state conditions was modelled (Neher & McGrath, 1957; IEC, 1994; IEEE, 1994) to account for the heat balance between the power dissipated in the conductor by the Joule effect, and the heat dissipated in the environment through the thermal resistance, RT (mKW-1), of the insulation and the soil, due to the temperature difference, ΔT (K), as shown in Eq.21. The electrical current rating, I (A), may then be calculated, as shown in Eq.22. TE2 RTRI  (21) TE RRTI (22) Refinements incorporating dielectric losses, qd (Wm-1), the number of conductors in the cable, n, eddy currents and circulating currents in metallic sheaths, (λ1,2), electrical resistance variation with conductor temperature, RE(Tc) (Ω), skin and proximity effects and the thermal resistance of each insulating layer, RT,1-4, of lead to the more complex expression for calculating the cable ampacity as given in Eq.23.   )]RR)(1(nR)1(nR[TR)]RRR(nR21[qTI4,T3,T212,T11,TcE4,T3,T2,T1,Td (23) Thermal resistances, RT,1-3, for cylindrical layers are calculated using Eq.24 and soil thermal resistance, RT,4, is modelled using Eq.25    ddD21ln2R Ts3,2,1T (24)   1Dz2Dz2ln2R2bbTs4T (25) where ρs-T represents the soil thermal resistivity (Wm-1K-1), D and d represent the respective external and internal diameters (m) for the calculated layer and zb represents the burial depth of the cable (m). Other calculation methods (IEC, 1994) have to be utilised when operating conditions differ from those stated above (for example when the cable is in a duct or in open air). The model described above requires detailed knowledge of the electric cable installation. However, this information may not always be available and therefore it is difficult to make practical use of the model. In these circumstances an alternative model given in Eq.26 may be used (ENA, 2004). The rated current of electric cables, I0 (A), is given in tables depending on the nominal voltage level, V (kV), the standardised cable cross-sectional area, A (m2) and laying conditions (trefoil, flat formation; in air, in ducts or directly buried). The dependence of the cable ampacity on the actual soil temperature, Ts (K) away from the rated soil temperature Ts rated (K) as well as the actual soil thermal resistivity away from the rated soil thermal resistivity is made linear through the coefficients ξT and ξρ respectively.        rated T,sT,srated ssT0 TTlaying,V,AII   (26) Since this research concerns the influence of environmental conditions on component ratings, the effect of the nominal voltage level which influences the dielectric loss, qd in Eq.23 is not considered. The effect of the heating given by adjacent components is also neglected as it is assumed that each cable has already been de-rated to take this effect into account. 4.3 Power transformers The thermal model (IEC, 2008) can be used to calculate the winding hot spot temperature for power transformers. This is the most important parameter since hotspot temperature exceedance can damage the transformer in two ways. Firstly, a temperature exceedance of 120ºC-140ºC can induce the formation of bubbles in the coolant oil, which in turn is liable to cause an insulation breakdown due to the local reduction of dielectric insulation strength. Secondly, high temperatures increase the ageing rate of the winding insulation. For this reason the maximum operating temperature should not exceed the rated value. The thermal model consists of a heat balance between the power dissipated in the winding and iron core, and the heat transferred to the environment via the refrigerating circuit. Considering the ambient temperature, Ta (K), thermal resistance between the windings and the oil, RT,W (mKW-1), the thermal resistance between the heat exchanger and the air, RT,HE (mKW-1) and the power dissipated into the core, I2RE, windings (W), it is possible to calculate the hot spot temperature, THS (K) as in Eq.27. Increasing the energy yield of generation from new and renewable energy resources 47Renewable Energy48  HE,TW,Twindings,E2aHS RRRITT  (27) Eq.27 is discussed (Susa et al., 2005) and leads to the IEC standard model for rating oil-filled power transformers as shown in Eq.28     yTOHSx2aTOaHS KTTr1rK1TTTT  (28) where TTO represents the top oil temperature (K), r represents the ratio between winding and core losses, K represents the loading ratio of the transformer, x represents the transformer oil exponent and y represents the transformer winding exponent. The maximum rating can be obtained by iteration, once the hot spot temperature has been set, and tabulated values for the parameters are given (IEC, 2008) for transformers with different types cooling system. Correction factors (IEC, 2008) can be used to model other operating conditions such as transformers operating within enclosures. Transformer cooling systems are classified with an acronym summarising (i) the coolant fluid: oil (O) or air (A), (ii) the convection around the core: natural (N), forced (F) or direct (D), (iii) the external refrigerating fluid: air (A) or water (W) and (iv) the external convection method: natural (N) or forced (F). Typically distribution transformers have ONAN or ONAF cooling systems. 5. Thermal state estimation This section describes the approach adopted to estimate, correct and interpolate environmental conditions to represent more accurately the actual environmental operating conditions for areas of the distribution network. This may then be used together with the models described in Section 4 to provide an estimation of power system component real-time thermal ratings. 5.1 Environmental condition estimations The inverse distance interpolation technique (Shepard, 1968) allows environmental conditions to be determined over a wide geographical area using a reduced set of inputs. This is attractive for situations where a large amount of installed measurements may be financially unattractive to the DNO or generator developer. The technique is also computationally efficient and allows the input locations to be readily adapted. The wind speed correction process is described below, as is the soil parameter correction process. Wind direction, air temperature and solar radiation values are included within interpolations but do not require the application of a correction factor. At each point in the geographical area, κ, the value of the parameter, Z, representing the environmental condition can be estimated as a weighted average of the parameter values known at ι points. The weighting factor is a function of the distance between the points as shown in Eq.29. 2,2,d1d1 ZZ (29) Ground roughness influences wind speed profiles and may lead to differences between the wind speed recorded by anemometers and the actual wind speed passing across an overhead line, particularly if the anemometer and overhead line are installed at different heights. This may be corrected using the wind profile power law given in Eq.30. The wind speed at two different heights is linked with the ground roughness through the exponent Kshear. Values of Kshear for different ground types are given in (IEC, 1991). ca KshearrefcKsheararefa zzzzWsWs  (30) Using Eq.30, the anemometer wind speed, Wsa (ms-1) at the weather station height, za (m) is extrapolated to a reference height, zref (m), to remove ground roughness dependence represented by the parameter Ksheara. The values from different anemometer locations may then be interpolated, using Eq.29 to provide a wind speed estimate at the reference height for a particular geographical location. The ground roughness at this location is then taken into account through the coefficient Kshearc along with the conductor height, zc (m) in Eq.30 to estimate the wind speed, Ws (ms-1) across the overhead line. Electric cable ratings are dependent on soil temperature and soil thermal resistivity, as well as cable construction, burial layout and burial depth (which is typically 0.8–1m). UK MetOffice datasets contain information regarding soil temperatures at a depth of 0.3m. However, no information is currently available from this source regarding soil thermal resistivity. Depth-dependent soil temperature distributions may be calculated using the Fourier law (Nairen et al., 2004) as shown in Eq.31      dzdTdzddtdT sTss (31) where Ts represents the soil temperature (K), t represents time (s), z represents soil depth (m) and δs-T(θ) represents the soil thermal diffusivity (m2s-1) as a function of gravimetric water content. Boundary conditions are set up for the lower soil layer (for example a constant temperature of 10oC at a depth of 2m) and soil temperature readings are used for the upper layer. Soil thermal resistivity ρs-T (mKW-1) may be calculated from Eq.32 using the soil thermal diffusivity, δs-T (m2s-1), the dry soil density ρs-density (kgm-3), and the soil thermal capacitance, Cs-T (Jkg-1K-1).   1TsdensitysTsTs C   (32) Soil thermal diffusivity and soil thermal capacity are influenced by soil composition, N (%), and soil gravimetric water content, θ and can be calculated using Eq.33-34 (Nidal, 2003). Increasing the energy yield of generation from new and renewable energy resources 49Renewable Energy48  HE,TW,Twindings,E2aHS RRRITT  (27) Eq.27 is discussed (Susa et al., 2005) and leads to the IEC standard model for rating oil-filled power transformers as shown in Eq.28     yTOHSx2aTOaHS KTTr1rK1TTTT  (28) where TTO represents the top oil temperature (K), r represents the ratio between winding and core losses, K represents the loading ratio of the transformer, x represents the transformer oil exponent and y represents the transformer winding exponent. The maximum rating can be obtained by iteration, once the hot spot temperature has been set, and tabulated values for the parameters are given (IEC, 2008) for transformers with different types cooling system. Correction factors (IEC, 2008) can be used to model other operating conditions such as transformers operating within enclosures. Transformer cooling systems are classified with an acronym summarising (i) the coolant fluid: oil (O) or air (A), (ii) the convection around the core: natural (N), forced (F) or direct (D), (iii) the external refrigerating fluid: air (A) or water (W) and (iv) the external convection method: natural (N) or forced (F). Typically distribution transformers have ONAN or ONAF cooling systems. 5. Thermal state estimation This section describes the approach adopted to estimate, correct and interpolate environmental conditions to represent more accurately the actual environmental operating conditions for areas of the distribution network. This may then be used together with the models described in Section 4 to provide an estimation of power system component real-time thermal ratings. 5.1 Environmental condition estimations The inverse distance interpolation technique (Shepard, 1968) allows environmental conditions to be determined over a wide geographical area using a reduced set of inputs. This is attractive for situations where a large amount of installed measurements may be financially unattractive to the DNO or generator developer. The technique is also computationally efficient and allows the input locations to be readily adapted. The wind speed correction process is described below, as is the soil parameter correction process. Wind direction, air temperature and solar radiation values are included within interpolations but do not require the application of a correction factor. At each point in the geographical area, κ, the value of the parameter, Z, representing the environmental condition can be estimated as a weighted average of the parameter values known at ι points. The weighting factor is a function of the distance between the points as shown in Eq.29. 2,2,d1d1 ZZ (29) Ground roughness influences wind speed profiles and may lead to differences between the wind speed recorded by anemometers and the actual wind speed passing across an overhead line, particularly if the anemometer and overhead line are installed at different heights. This may be corrected using the wind profile power law given in Eq.30. The wind speed at two different heights is linked with the ground roughness through the exponent Kshear. Values of Kshear for different ground types are given in (IEC, 1991). ca KshearrefcKsheararefa zzzzWsWs  (30) Using Eq.30, the anemometer wind speed, Wsa (ms-1) at the weather station height, za (m) is extrapolated to a reference height, zref (m), to remove ground roughness dependence represented by the parameter Ksheara. The values from different anemometer locations may then be interpolated, using Eq.29 to provide a wind speed estimate at the reference height for a particular geographical location. The ground roughness at this location is then taken into account through the coefficient Kshearc along with the conductor height, zc (m) in Eq.30 to estimate the wind speed, Ws (ms-1) across the overhead line. Electric cable ratings are dependent on soil temperature and soil thermal resistivity, as well as cable construction, burial layout and burial depth (which is typically 0.8–1m). UK MetOffice datasets contain information regarding soil temperatures at a depth of 0.3m. However, no information is currently available from this source regarding soil thermal resistivity. Depth-dependent soil temperature distributions may be calculated using the Fourier law (Nairen et al., 2004) as shown in Eq.31      dzdTdzddtdT sTss (31) where Ts represents the soil temperature (K), t represents time (s), z represents soil depth (m) and δs-T(θ) represents the soil thermal diffusivity (m2s-1) as a function of gravimetric water content. Boundary conditions are set up for the lower soil layer (for example a constant temperature of 10oC at a depth of 2m) and soil temperature readings are used for the upper layer. Soil thermal resistivity ρs-T (mKW-1) may be calculated from Eq.32 using the soil thermal diffusivity, δs-T (m2s-1), the dry soil density ρs-density (kgm-3), and the soil thermal capacitance, Cs-T (Jkg-1K-1).   1TsdensitysTsTs C   (32) Soil thermal diffusivity and soil thermal capacity are influenced by soil composition, N (%), and soil gravimetric water content, θ and can be calculated using Eq.33-34 (Nidal, 2003). Increasing the energy yield of generation from new and renewable energy resources 49Renewable Energy50     79.4N209.08.14Ts (33)   81.5753.0N00561.0224.0C densitysTs (34) Ground water content may be determined using the closed form of Richard’s equation (Celia et al., 1990) as described in Eq.35 after the calculation of the unsaturated hydraulic diffusivity, δs-θ(θ) (m2s-1) and the unsaturated hydraulic conductivity, ks-θ(θ) (ms-1) (Van Genuchten, 1980).      ss kdzddzddtd (35) In order to solve Eq.35, boundary and initial conditions must be specified. A constant water content equal to the saturation value can be set at a depth of 2.5 metres, corresponding to the UK water table. Furthermore, the ground-level water content can be linked to MetOffice rainfall values, lr (mm) using the model described in Eq.36, where Krain1 represents the normalized soil water loss (day-1) and Krain2 represents the normalized net rainfall coefficient (day-1mm-1) (Rodriguez-Iturbe, 1996). )t(lKrain)t(Kraindtdr21  (36) 5.2 Component rating estimations In this section a description of the algorithm responsible for the state estimation is given. The primary aim of the thermal state estimation algorithm is to allow the rating of components, which are not directly monitored within the power system, to be assessed. Thermal state estimations facilitate the precise and reliable assessment of environmental conditions whereby a minimal amount of meteorological monitoring installations facilitate the assessment of component thermal ratings within a wide area. This may then be validated through the carefully selected monitoring of component operating temperatures. The algorithm provides a reliable estimation of power system component thermal ratings described by an appropriate cumulative probability function. A state estimation technique based on the Montecarlo method is used, giving a more complete description of the possible states of the system. The minimum, maximum, average and standard deviation of component rating forecasts may be calculated according to the possible forecasted weather conditions. As necessary for overhead lines and electric cables, each component is divided into sections to take into account different thermal operating conditions such as overhead line orientations and changes in electric cable installation conditions. The section resulting in the lowest rating values is then used to provide a rating for the entire component. Furthermore, the deployment of a real-time thermal rating system underpinned by thermal state estimation techniques has the potential to reduce the necessity of auxiliary communications infrastructure whilst simultaneously increasing the reliability of the system if measurement or communication failures occur. Fig. 2. Component rating estimation flow chart The thermal state estimation algorithm is illustrated in Figure 2, where it is possible to see the following steps: 1. Forecasted weather data is read from an external source (in this case the database “a”), where real-time readings from different meteorological station locations within the distribution network are stored. 2. A set of values for weather parameters is calculated in the following way: From the data read in “1” the parameters of a cumulative probability function are calculated. In this case the Beta probability function is used. A random value for the probability is selected and from the cumulative probability function the corresponding parameter value is found. This is repeated for each weather parameter. 3. For each section of the component the values of local environmental condition are calculated with the methodology described in Section 5.1, then the rating is calculated using the models described in Section 4. The result is stored temporarily in “b”. 4. The component rating is calculated by selecting the minimum rating of each section. The results are temporarily stored in “c”. 5. The steps from 2 to 4 are repeated for a fixed number of times γ. 6. The precision of the result is compared with a predefined value. If the result is not acceptable, a new value for γ is calculated and the steps from 2 to 5 are repeated 7. Component ratings stored in “c” are analysed in order to calculate the minimum, maximum, average value and standard deviation for each component of the network. 6. Generator output control strategies This section describes strategies for the power output control of single and multiple generation installations. Increasing the energy yield of generation from new and renewable energy resources 51Renewable Energy50     79.4N209.08.14Ts (33)   81.5753.0N00561.0224.0C densitysTs (34) Ground water content may be determined using the closed form of Richard’s equation (Celia et al., 1990) as described in Eq.35 after the calculation of the unsaturated hydraulic diffusivity, δs-θ(θ) (m2s-1) and the unsaturated hydraulic conductivity, ks-θ(θ) (ms-1) (Van Genuchten, 1980).      ss kdzddzddtd (35) In order to solve Eq.35, boundary and initial conditions must be specified. A constant water content equal to the saturation value can be set at a depth of 2.5 metres, corresponding to the UK water table. Furthermore, the ground-level water content can be linked to MetOffice rainfall values, lr (mm) using the model described in Eq.36, where Krain1 represents the normalized soil water loss (day-1) and Krain2 represents the normalized net rainfall coefficient (day-1mm-1) (Rodriguez-Iturbe, 1996). )t(lKrain)t(Kraindtdr21  (36) 5.2 Component rating estimations In this section a description of the algorithm responsible for the state estimation is given. The primary aim of the thermal state estimation algorithm is to allow the rating of components, which are not directly monitored within the power system, to be assessed. Thermal state estimations facilitate the precise and reliable assessment of environmental conditions whereby a minimal amount of meteorological monitoring installations facilitate the assessment of component thermal ratings within a wide area. This may then be validated through the carefully selected monitoring of component operating temperatures. The algorithm provides a reliable estimation of power system component thermal ratings described by an appropriate cumulative probability function. A state estimation technique based on the Montecarlo method is used, giving a more complete description of the possible states of the system. The minimum, maximum, average and standard deviation of component rating forecasts may be calculated according to the possible forecasted weather conditions. As necessary for overhead lines and electric cables, each component is divided into sections to take into account different thermal operating conditions such as overhead line orientations and changes in electric cable installation conditions. The section resulting in the lowest rating values is then used to provide a rating for the entire component. Furthermore, the deployment of a real-time thermal rating system underpinned by thermal state estimation techniques has the potential to reduce the necessity of auxiliary communications infrastructure whilst simultaneously increasing the reliability of the system if measurement or communication failures occur. Fig. 2. Component rating estimation flow chart The thermal state estimation algorithm is illustrated in Figure 2, where it is possible to see the following steps: 1. Forecasted weather data is read from an external source (in this case the database “a”), where real-time readings from different meteorological station locations within the distribution network are stored. 2. A set of values for weather parameters is calculated in the following way: From the data read in “1” the parameters of a cumulative probability function are calculated. In this case the Beta probability function is used. A random value for the probability is selected and from the cumulative probability function the corresponding parameter value is found. This is repeated for each weather parameter. 3. For each section of the component the values of local environmental condition are calculated with the methodology described in Section 5.1, then the rating is calculated using the models described in Section 4. The result is stored temporarily in “b”. 4. The component rating is calculated by selecting the minimum rating of each section. The results are temporarily stored in “c”. 5. The steps from 2 to 4 are repeated for a fixed number of times γ. 6. The precision of the result is compared with a predefined value. If the result is not acceptable, a new value for γ is calculated and the steps from 2 to 5 are repeated 7. Component ratings stored in “c” are analysed in order to calculate the minimum, maximum, average value and standard deviation for each component of the network. 6. Generator output control strategies This section describes strategies for the power output control of single and multiple generation installations. Increasing the energy yield of generation from new and renewable energy resources 51Renewable Energy52 6.1 Single generation installations A number of different solutions are presented (ENA, 2003) that may be developed to manage, in an active manner, the power flows associated with the connection of a single generation installation. The most basic systems involve the disconnection of generation in the event that network power flow congestion occurs when utilising component static thermal ratings. This solution may be developed further by actively switching between seasonal ratings based on fixed meteorological assumptions and adjusting the number of disconnected generation units accordingly. More sophisticated solutions are developed from the principle of generator power output control, utilising technologies such as the pitch control of wind turbine blades to capture a desired amount of wind energy. In this approach the powers flowing in the critical feeders of the network are monitored, taking load demand into account, and the power exported from the generator is controlled to ensure the capability of the network is not exceeded. This may be developed as a solution utilising component static ratings, with demand-following control of the generator output, or as a solution utilising component real-time thermal ratings, with demand-following control of the generator output. Although generators may, at times, be requested to operate in voltage control mode, it is in the interest of the operators to maximise annual exported active energy (MWh) as that directly relates to the revenue of the generation installation owner. Thus it is assumed that the control system outlined in this chapter is utilised for generators exporting real power at unity power factor. An assessment of the real power output adjustment required by a generator may be calculated in Eq. 37-38, relating to the component power flow sensitivity factor and the real-time thermal rating of the component  m,Pk,ik,im,PdGdPPG (37)           2k,i2k,i2k,i2limTark,i Q'S'Q''SUP (38) where ΔGP,m (MW) is the required change in real power output of the generator at node m, ΔPi,k (MW) is the required change in real power flow through the component, from node i to node k in order to bring the resultant power flow back within thermal limits, dPi,k/dGP,m is the power flow sensitivity factor that relates the change in nodal real power injection at m with the change in real power flow seen in the component from node i to node k, ’Si,k (MVA) is the apparent power flow in the component before control actions are implemented, ’Qi,k (MVAr) is the reactive power flow in the component from node i to node k before the control actions are implemented, UTar is the target utilisation limit of the component after the control actions have been implemented, Slim (MVA) is the static or real-time thermal rating of the component and “Qi,k (MVAr) is the reactive power flow in the component from node i to node k after the control actions have been implemented. Rather than neglecting MVAr flow, it may be assumed constant for a particular operating condition. Thus dPi,k/dGP,m >> dQi,k/dGP,m and a simplification can be made to Eq.38 that “Qi,k ≈ ’Qi,k. 6.2 Multiple generation installations Present last-in first-off control strategies include the complete disconnection of generation installations or the power output reduction of generators in discrete intervals (Roberts, 2004). A step beyond this could be to implement a last-in first-off ’sensitivity-based’ control strategy whereby generator power outputs are adjusted (in contractual order) based on Eq.37-38. Moreover, these equations may be developed further to define the power output adjustment of multiple generation installations (assuming contracts are in place to allow generator operation outside of a last-in first-off constraint priority) in order to maximise the annual energy yield of the aggregated generation installations. Two candidate strategies are outlined below: 1. Equal percentage reduction of generator power output The equal percentage reduction signal, Φ, broadcast to multiple generators in order to proportionally reduce their power output (taking power flow sensitivity factors into account) may be assessed as in Eq.39-40.  m,Pm,P GG (39)    m0mm,Pikm,PikdGdPGP (40) This strategy may be seen as the most ‘fair’ option since each generator is constrained as an equal proportion of their present power output. 2. Most appropriate technical strategy In order to minimise the overall generator constraint and thus maximise the annual energy yield of the aggregated generators, an assessment of the required generator power output adjustment is given in Eq.41. In this case, rather than adjusting generators in contractual order according to the historical order of connection agreements, the generator constraint order is prioritised according to the magnitude of the power flow sensitivity factors. Thus the generator with the greatest technical ability to solve the power flow congestion is adjusted first.  maxm,Pk,iikminm,P }dGdP{PG (41) 7. Case study: Power output control of multiple generators informed by component thermal ratings The following case study illustrates various aspects of the development of a power output control system for generation installations. An assessment of power flow sensitivity factors is made which allows thermal vulnerability factors to be calculated. This informs the Increasing the energy yield of generation from new and renewable energy resources 53Renewable Energy52 6.1 Single generation installations A number of different solutions are presented (ENA, 2003) that may be developed to manage, in an active manner, the power flows associated with the connection of a single generation installation. The most basic systems involve the disconnection of generation in the event that network power flow congestion occurs when utilising component static thermal ratings. This solution may be developed further by actively switching between seasonal ratings based on fixed meteorological assumptions and adjusting the number of disconnected generation units accordingly. More sophisticated solutions are developed from the principle of generator power output control, utilising technologies such as the pitch control of wind turbine blades to capture a desired amount of wind energy. In this approach the powers flowing in the critical feeders of the network are monitored, taking load demand into account, and the power exported from the generator is controlled to ensure the capability of the network is not exceeded. This may be developed as a solution utilising component static ratings, with demand-following control of the generator output, or as a solution utilising component real-time thermal ratings, with demand-following control of the generator output. Although generators may, at times, be requested to operate in voltage control mode, it is in the interest of the operators to maximise annual exported active energy (MWh) as that directly relates to the revenue of the generation installation owner. Thus it is assumed that the control system outlined in this chapter is utilised for generators exporting real power at unity power factor. An assessment of the real power output adjustment required by a generator may be calculated in Eq. 37-38, relating to the component power flow sensitivity factor and the real-time thermal rating of the component  m,Pk,ik,im,PdGdPPG (37)           2k,i2k,i2k,i2limTark,i Q'S'Q''SUP (38) where ΔGP,m (MW) is the required change in real power output of the generator at node m, ΔPi,k (MW) is the required change in real power flow through the component, from node i to node k in order to bring the resultant power flow back within thermal limits, dPi,k/dGP,m is the power flow sensitivity factor that relates the change in nodal real power injection at m with the change in real power flow seen in the component from node i to node k, ’Si,k (MVA) is the apparent power flow in the component before control actions are implemented, ’Qi,k (MVAr) is the reactive power flow in the component from node i to node k before the control actions are implemented, UTar is the target utilisation limit of the component after the control actions have been implemented, Slim (MVA) is the static or real-time thermal rating of the component and “Qi,k (MVAr) is the reactive power flow in the component from node i to node k after the control actions have been implemented. Rather than neglecting MVAr flow, it may be assumed constant for a particular operating condition. Thus dPi,k/dGP,m >> dQi,k/dGP,m and a simplification can be made to Eq.38 that “Qi,k ≈ ’Qi,k. 6.2 Multiple generation installations Present last-in first-off control strategies include the complete disconnection of generation installations or the power output reduction of generators in discrete intervals (Roberts, 2004). A step beyond this could be to implement a last-in first-off ’sensitivity-based’ control strategy whereby generator power outputs are adjusted (in contractual order) based on Eq.37-38. Moreover, these equations may be developed further to define the power output adjustment of multiple generation installations (assuming contracts are in place to allow generator operation outside of a last-in first-off constraint priority) in order to maximise the annual energy yield of the aggregated generation installations. Two candidate strategies are outlined below: 1. Equal percentage reduction of generator power output The equal percentage reduction signal, Φ, broadcast to multiple generators in order to proportionally reduce their power output (taking power flow sensitivity factors into account) may be assessed as in Eq.39-40.  m,Pm,P GG (39)    m0mm,Pikm,PikdGdPGP (40) This strategy may be seen as the most ‘fair’ option since each generator is constrained as an equal proportion of their present power output. 2. Most appropriate technical strategy In order to minimise the overall generator constraint and thus maximise the annual energy yield of the aggregated generators, an assessment of the required generator power output adjustment is given in Eq.41. In this case, rather than adjusting generators in contractual order according to the historical order of connection agreements, the generator constraint order is prioritised according to the magnitude of the power flow sensitivity factors. Thus the generator with the greatest technical ability to solve the power flow congestion is adjusted first.  maxm,Pk,iikminm,P }dGdP{PG (41) 7. Case study: Power output control of multiple generators informed by component thermal ratings The following case study illustrates various aspects of the development of a power output control system for generation installations. An assessment of power flow sensitivity factors is made which allows thermal vulnerability factors to be calculated. This informs the Increasing the energy yield of generation from new and renewable energy resources 53Renewable Energy54 development of a real-time thermal rating algorithm for an electric cable section within the case study network. An offline simulation is conducted which allows the energy yield of candidate strategies for the power output control of multiple generators (informed by static and real-time thermal ratings) to be assessed. 7.1 Case study network description The case study network shown in Figure 3 was derived from a section of the United Kingdom generic distribution system (UKGDS) ‘EHV3’ network (Foote et al., 2006). A hydro generator and two combined heat and power (CHP) generators were connected to the network at 33kV nodes. A summary of the generator properties is given in Table 1. Generator node, m Generator type Installed capacity (MW) Last-in first-off constraint order dP353,357/dGP,m 352 Hydro 18 2 0.475 353 CHP 40 3 0.477 354 CHP 30 1 0.327 Table 1. Generator properties Fig. 3. United Kingdom generic distribution system case study 7.2 Thermal vulnerability factor assessment The most thermally vulnerable components within the case study network were identified using a thermal vulnerability factor assessment as given in Figure 4 (and detailed in Sections 3.1 and 3.2). From this assessment it may be seen that a generator real power injection at node 352 increases the thermal vulnerability of the components between nodes 352 and 353, and between nodes 353 and 357 but reduces the thermal vulnerability of the component between nodes 316 and 353. Moreover, a generator real power injection at node 353 increases the thermal vulnerability of the component between nodes 353 and 357 but reduces the thermal vulnerability of the component between nodes 316 and 353. Furthermore, a generator real power injection at node 354 increases the thermal vulnerability of the components between nodes 316 and 353, 354 and 316, and 353 and 357. However, this injection has little impact on the thermal vulnerability of the component between nodes 352 and 353. In this particular case, multiple generator injections cause the thermal vulnerability factors to accumulate for the electric cable between nodes 353 and 357, making it the most thermally vulnerable component. Having identified this component as being a potential thermal pinch-point and causing power flow congestion within the network, the targeted development of a real-time thermal rating system is informed. 7.3 Component real-time thermal rating assessment A summary of the characteristics of the thermally vulnerable electric cable are given in Table 2. The ’Valley’ UK MetOffice dataset for the calendar year 2006 was used with the thermal state estimation technique described by Eq.31-36 (Section 5.1) to estimate environmental operating conditions local to the electric cable. This information was used to populate the electric cable steady-state thermal model described by Eq.26 (Section 4.2). The resulting real-time thermal rating for the electric cable, together with the static rating, is given in Figure 5. It was found that the average uprating for the electric cable, based on minimum daily ratings was 6.0% across the year. Generator injection node:-4-3-2-10123456316 - 353 354 - 316 352 - 353 353 - 357Component (From node - to node)Thermal vulnerability factor (TVF)352 353 354 Fig. 4. Thermally vulnerable component assessment for UKGDS case study Increasing the energy yield of generation from new and renewable energy resources 55Thermal vulnerability factor (TVF)Renewable Energy54 development of a real-time thermal rating algorithm for an electric cable section within the case study network. An offline simulation is conducted which allows the energy yield of candidate strategies for the power output control of multiple generators (informed by static and real-time thermal ratings) to be assessed. 7.1 Case study network description The case study network shown in Figure 3 was derived from a section of the United Kingdom generic distribution system (UKGDS) ‘EHV3’ network (Foote et al., 2006). A hydro generator and two combined heat and power (CHP) generators were connected to the network at 33kV nodes. A summary of the generator properties is given in Table 1. Generator node, m Generator type Installed capacity (MW) Last-in first-off constraint order dP353,357/dGP,m 352 Hydro 18 2 0.475 353 CHP 40 3 0.477 354 CHP 30 1 0.327 Table 1. Generator properties Fig. 3. United Kingdom generic distribution system case study 7.2 Thermal vulnerability factor assessment The most thermally vulnerable components within the case study network were identified using a thermal vulnerability factor assessment as given in Figure 4 (and detailed in Sections 3.1 and 3.2). From this assessment it may be seen that a generator real power injection at node 352 increases the thermal vulnerability of the components between nodes 352 and 353, and between nodes 353 and 357 but reduces the thermal vulnerability of the component between nodes 316 and 353. Moreover, a generator real power injection at node 353 increases the thermal vulnerability of the component between nodes 353 and 357 but reduces the thermal vulnerability of the component between nodes 316 and 353. Furthermore, a generator real power injection at node 354 increases the thermal vulnerability of the components between nodes 316 and 353, 354 and 316, and 353 and 357. However, this injection has little impact on the thermal vulnerability of the component between nodes 352 and 353. In this particular case, multiple generator injections cause the thermal vulnerability factors to accumulate for the electric cable between nodes 353 and 357, making it the most thermally vulnerable component. Having identified this component as being a potential thermal pinch-point and causing power flow congestion within the network, the targeted development of a real-time thermal rating system is informed. 7.3 Component real-time thermal rating assessment A summary of the characteristics of the thermally vulnerable electric cable are given in Table 2. The ’Valley’ UK MetOffice dataset for the calendar year 2006 was used with the thermal state estimation technique described by Eq.31-36 (Section 5.1) to estimate environmental operating conditions local to the electric cable. This information was used to populate the electric cable steady-state thermal model described by Eq.26 (Section 4.2). The resulting real-time thermal rating for the electric cable, together with the static rating, is given in Figure 5. It was found that the average uprating for the electric cable, based on minimum daily ratings was 6.0% across the year. Generator injection node:-4-3-2-10123456316 - 353 354 - 316 352 - 353 353 - 357Component (From node - to node)Thermal vulnerability factor (TVF)352 353 354 Fig. 4. Thermally vulnerable component assessment for UKGDS case study Increasing the energy yield of generation from new and renewable energy resources 55Thermal vulnerability factor (TVF)Renewable Energy56 Electric cable characteristic Value / category Resistance, RE (pu) 0.01489 Reactance, X (pu) 0.01296 Cross-sectional area, A (mm2) 150 Standard thermal rating, Slim (MVA) 30 Nominal line-to-line voltage, V (kV) 33 Standardised phase current rating, I0 (A) 525 Installation conditions Flat formation Buried in a duct Table 2. Electric cable characteristics 7.4 Electro-thermal simulation approach This section describes the offline electro-thermal simulation that was used to quantify the individual and aggregated generator annual energy yields (as seen in Section 7.6) for the candidate multiple generator control strategies. UKGDS annual ½ hour loading and generation profiles were utilised for the simulation of network power flows and busbar voltages through a load flow software tool. The open-loop power output control system compared network power flows to component static and real-time thermal limits for each ½ hour interval within the simulated year. When power flow congestion was detected (signified in this case by an electric cable utilisation, U353,357 > 1) the controller implemented the relevant generator power output control algorithm as described below. A validating load flow simulation was conducted as part of the control solution to ensure that the updated generator power outputs solved the power flow congestion and did not breach busbar voltage limits (set to ±6% of nominal). 2829303132333435Jan Apr Jul OctTime (Months)Rating (MVA)Static thermal ratingReal-time thermal rating Fig. 5. Electric cable static and real-time thermal rating The individual and aggregated generator power outputs were integrated across the simulated year in order to calculate the potential annual energy yields resulting from the candidate multiple generator control strategies, informed by component static and real-time thermal limits. The results are summarised in Section 7.6. 1. Last-in first-off discrete interval generator output reduction In order to solve the power flow congestion, the power output of the generators was reduced in discrete intervals of 33% according to the contractual priority order given in Table 1. Thus the power output of the generator at node 354 was reduced by 33% then 66% followed by generator disconnection and the 33% power output reduction of the generator at node 352. 2. Last-in first off sensitivity-based generator output reduction Using Eq.38 with UTar = 0.99, the required real power reduction in the electric cable, ΔP353,357 in order to bring the resultant power flow back within the desired thermal limit was calculated. The required reduction in the power output of the first generator (contractually) to be constrained, ΔGP,354, was calculated using Eq.42 for GP,354 – ΔGP,354>0 and the output from GP,353 and GP,352 was unconstrained. On occasions when the required constraint of the generator at node 354 would result in GP,354 – ΔGP,354<0, the generator was disconnected and the remaining power flow congestion was solved by the next generator (at node 352) to be contractually constrained. 3270357353354 .PG ,,P (42) 3. Equal percentage reduction of generator power outputs Using Eq.38 with UTar = 0.99, the required real power reduction in the electric cable, ΔP353,357 in order to bring the resultant power flow back within the desired thermal limit was calculated. The equal percentage reduction signal, Φ, broadcast to all generators in order to solve the power flow congestion was calculated and implemented using Eq.43-46. 354353352357353327047704750 ,P,P,P,G.G.G.P (43)  352352 ,P,P GG (44)  353353 ,P,P GG (45)  354354 ,P,P GG (46) Increasing the energy yield of generation from new and renewable energy resources 57Rating (MVA)Renewable Energy56 Electric cable characteristic Value / category Resistance, RE (pu) 0.01489 Reactance, X (pu) 0.01296 Cross-sectional area, A (mm2) 150 Standard thermal rating, Slim (MVA) 30 Nominal line-to-line voltage, V (kV) 33 Standardised phase current rating, I0 (A) 525 Installation conditions Flat formation Buried in a duct Table 2. Electric cable characteristics 7.4 Electro-thermal simulation approach This section describes the offline electro-thermal simulation that was used to quantify the individual and aggregated generator annual energy yields (as seen in Section 7.6) for the candidate multiple generator control strategies. UKGDS annual ½ hour loading and generation profiles were utilised for the simulation of network power flows and busbar voltages through a load flow software tool. The open-loop power output control system compared network power flows to component static and real-time thermal limits for each ½ hour interval within the simulated year. When power flow congestion was detected (signified in this case by an electric cable utilisation, U353,357 > 1) the controller implemented the relevant generator power output control algorithm as described below. A validating load flow simulation was conducted as part of the control solution to ensure that the updated generator power outputs solved the power flow congestion and did not breach busbar voltage limits (set to ±6% of nominal). 2829303132333435Jan Apr Jul OctTime (Months)Rating (MVA)Static thermal ratingReal-time thermal rating Fig. 5. Electric cable static and real-time thermal rating The individual and aggregated generator power outputs were integrated across the simulated year in order to calculate the potential annual energy yields resulting from the candidate multiple generator control strategies, informed by component static and real-time thermal limits. The results are summarised in Section 7.6. 1. Last-in first-off discrete interval generator output reduction In order to solve the power flow congestion, the power output of the generators was reduced in discrete intervals of 33% according to the contractual priority order given in Table 1. Thus the power output of the generator at node 354 was reduced by 33% then 66% followed by generator disconnection and the 33% power output reduction of the generator at node 352. 2. Last-in first off sensitivity-based generator output reduction Using Eq.38 with UTar = 0.99, the required real power reduction in the electric cable, ΔP353,357 in order to bring the resultant power flow back within the desired thermal limit was calculated. The required reduction in the power output of the first generator (contractually) to be constrained, ΔGP,354, was calculated using Eq.42 for GP,354 – ΔGP,354>0 and the output from GP,353 and GP,352 was unconstrained. On occasions when the required constraint of the generator at node 354 would result in GP,354 – ΔGP,354<0, the generator was disconnected and the remaining power flow congestion was solved by the next generator (at node 352) to be contractually constrained. 3270357353354 .PG ,,P (42) 3. Equal percentage reduction of generator power outputs Using Eq.38 with UTar = 0.99, the required real power reduction in the electric cable, ΔP353,357 in order to bring the resultant power flow back within the desired thermal limit was calculated. The equal percentage reduction signal, Φ, broadcast to all generators in order to solve the power flow congestion was calculated and implemented using Eq.43-46. 354353352357353327047704750 ,P,P,P,G.G.G.P (43)  352352 ,P,P GG (44)  353353 ,P,P GG (45)  354354 ,P,P GG (46) Increasing the energy yield of generation from new and renewable energy resources 57Rating (MVA)Renewable Energy58 4. Most appropriate technical strategy Using Eq.38 with UTar = 0.99, the required real power reduction in the electric cable, ΔP353,357 in order to bring the resultant power flow back within the desired thermal limit was calculated. Since, in this case, Gp,353 has the greatest technical ability to solve power flow congestion (as assessed by the comparative magnitude of power flow sensitivity factors in Table 1), Eq.47 was implemented to minimise the overall generation constraint and thus maximise the annual energy yield of the aggregated generators. On occasions when the required constraint of the generator at node 353 would result in GP,353 – ΔGP,353<0, the generator was disconnected and the remaining power flow congestion was solved by the generator with the next-greatest power flow sensitivity factor (in this case at node 352). 4770357353353 .PG ,,P (47) 7.5 Findings The marginal annual energy yields resulting from the applied candidate control strategies informed by the electric cable real-time thermal ratings are summarised in Table 3. A similar table is given (Jupe & Taylor, 2009a) for the marginal energy yields resulting from static component ratings. The last-in first-off ’discrete interval’ strategy with static component ratings was used as a datum with annual energy yields of 106.38GWh, 298.26GWh, 120.84 GWh and 525.48GWh for GP, 352, GP, 353, GP, 354 and GP, aggregated respectively. Generator control strategy Generation marginal annual energy yield (%) with electric cable real-time thermal rating GP, 352 GP, 353 GP, 354 GP, aggregated Last-in first-off (Discrete-interval) 1.7 0.0 25.8 6.3 Last-in first-off (Sensitivity-based) 1.7 0.0 41.9 10.0 Equal % reduction of power output -5.5 -6.2 73.6 12.3 Most appropriate technical strategy 1.7 -12.0 85.1 13.1 Table 3. Generation marginal annual energy yields 8. Discussion Considering Table 3, for the case study network it can be seen that the adoption of both real-time thermal rating systems and sensitivity factor-based generator power output control strategies have the potential to unlock gains in the aggregated annual energy yield contribution of multiple generation installations. Moving from a static rating system to a real-time rating system has the potential to unlock an extra 6.3% aggregated annual energy yield for the last-in first-off ’discrete interval’ power output reduction strategy. If a last-in first-off contractual priority is retained but a sensitivity-based power output control strategy is adopted this has the potential to unlock an extra 3.7% marginal aggregated annual energy yield. Moreover, a further 3.1% annual energy yield gain may be achieved by utilising the ‘most appropriate technical’ strategy that minimises overall generator constraint. For the ‘equal percentage reduction of generator power output’ strategy, it can be seen that the relative annual energy yields of GP,352 and GP,353 are reduced in order to achieve an aggregated generator annual energy yield gain. This phenomenon is even more pronounced in the ‘most appropriate technical’ strategy where annual energy yield gains of 1.7% and 85.1% for GP,352 and GP,354 respectively are facilitated by the 12.0% reduction in the annual energy yield of GP,353. The aim of this research was to outline the developmental stages required in the output control of generation installations for improved new and renewable energy yield penetrations. The purpose of the thermal vulnerability factor assessment was to identify thermally vulnerable components within distribution networks. It has been shown that the thermal vulnerability factor assessment is appropriate for use in a strategic way to assess longer term and wide-spread generation growth scenarios. Therefore the thermal vulnerability factor assessment procedure would be valuable for DNOs looking to develop long-term generation accommodation strategies for areas of their network. Furthermore, the thermal vulnerability factor assessment identifies those components that would most benefit from being thermally monitored to unlock latent power flow capacity through a real-time thermal rating system. The derived power flow sensitivity factors are network configuration-specific and assume that the network configuration will not be frequently changing. It is feasible, however, to develop an online control system that makes use of alternative sets of the above mentioned predetermined power flow sensitivity factors based on network switch status information. In order to deploy the control strategies presented there clearly needs to be a contractual mechanism in place that gives an incentive to separately owned generators to have their power output reduced in order to maximise the overall annual energy yield. One such mechanism could be to set up cross-payments between generators whereby the generators that are constrained to maximise the aggregated energy yield are rewarded by payments from the other generators. 9. Conclusion It is anticipated that distribution networks will continue to see a proliferation of generation from new and RE resources in coming years. In some cases this will result in power flow congestion with thermally vulnerable components restricting the installed capacity and energy yield penetration of generators. Therefore a system that can be developed for the management of power flows within distribution networks, through the power output control of generators, could be of great benefit. The development stages of such a system were outlined in Sections 3 – 6. These included: (i) the identification of thermally vulnerable components through a thermal vulnerability factor assessment that related power flow sensitivity factors to component thermal limits, (ii) the targeted development of thermal models to allow latent capacity in the network assets to be unlocked through a real-time thermal rating system, (iii) the development of thermal state estimation techniques that would allow component steady-state thermal models to be populated with real-time data from meteorological monitoring equipment installed in the distribution network (iv) the use Increasing the energy yield of generation from new and renewable energy resources 59Renewable Energy58 4. Most appropriate technical strategy Using Eq.38 with UTar = 0.99, the required real power reduction in the electric cable, ΔP353,357 in order to bring the resultant power flow back within the desired thermal limit was calculated. Since, in this case, Gp,353 has the greatest technical ability to solve power flow congestion (as assessed by the comparative magnitude of power flow sensitivity factors in Table 1), Eq.47 was implemented to minimise the overall generation constraint and thus maximise the annual energy yield of the aggregated generators. On occasions when the required constraint of the generator at node 353 would result in GP,353 – ΔGP,353<0, the generator was disconnected and the remaining power flow congestion was solved by the generator with the next-greatest power flow sensitivity factor (in this case at node 352). 4770357353353 .PG ,,P (47) 7.5 Findings The marginal annual energy yields resulting from the applied candidate control strategies informed by the electric cable real-time thermal ratings are summarised in Table 3. A similar table is given (Jupe & Taylor, 2009a) for the marginal energy yields resulting from static component ratings. The last-in first-off ’discrete interval’ strategy with static component ratings was used as a datum with annual energy yields of 106.38GWh, 298.26GWh, 120.84 GWh and 525.48GWh for GP, 352, GP, 353, GP, 354 and GP, aggregated respectively. Generator control strategy Generation marginal annual energy yield (%) with electric cable real-time thermal rating GP, 352 GP, 353 GP, 354 GP, aggregated Last-in first-off (Discrete-interval) 1.7 0.0 25.8 6.3 Last-in first-off (Sensitivity-based) 1.7 0.0 41.9 10.0 Equal % reduction of power output -5.5 -6.2 73.6 12.3 Most appropriate technical strategy 1.7 -12.0 85.1 13.1 Table 3. Generation marginal annual energy yields 8. Discussion Considering Table 3, for the case study network it can be seen that the adoption of both real-time thermal rating systems and sensitivity factor-based generator power output control strategies have the potential to unlock gains in the aggregated annual energy yield contribution of multiple generation installations. Moving from a static rating system to a real-time rating system has the potential to unlock an extra 6.3% aggregated annual energy yield for the last-in first-off ’discrete interval’ power output reduction strategy. If a last-in first-off contractual priority is retained but a sensitivity-based power output control strategy is adopted this has the potential to unlock an extra 3.7% marginal aggregated annual energy yield. Moreover, a further 3.1% annual energy yield gain may be achieved by utilising the ‘most appropriate technical’ strategy that minimises overall generator constraint. For the ‘equal percentage reduction of generator power output’ strategy, it can be seen that the relative annual energy yields of GP,352 and GP,353 are reduced in order to achieve an aggregated generator annual energy yield gain. This phenomenon is even more pronounced in the ‘most appropriate technical’ strategy where annual energy yield gains of 1.7% and 85.1% for GP,352 and GP,354 respectively are facilitated by the 12.0% reduction in the annual energy yield of GP,353. The aim of this research was to outline the developmental stages required in the output control of generation installations for improved new and renewable energy yield penetrations. The purpose of the thermal vulnerability factor assessment was to identify thermally vulnerable components within distribution networks. It has been shown that the thermal vulnerability factor assessment is appropriate for use in a strategic way to assess longer term and wide-spread generation growth scenarios. Therefore the thermal vulnerability factor assessment procedure would be valuable for DNOs looking to develop long-term generation accommodation strategies for areas of their network. Furthermore, the thermal vulnerability factor assessment identifies those components that would most benefit from being thermally monitored to unlock latent power flow capacity through a real-time thermal rating system. The derived power flow sensitivity factors are network configuration-specific and assume that the network configuration will not be frequently changing. It is feasible, however, to develop an online control system that makes use of alternative sets of the above mentioned predetermined power flow sensitivity factors based on network switch status information. In order to deploy the control strategies presented there clearly needs to be a contractual mechanism in place that gives an incentive to separately owned generators to have their power output reduced in order to maximise the overall annual energy yield. One such mechanism could be to set up cross-payments between generators whereby the generators that are constrained to maximise the aggregated energy yield are rewarded by payments from the other generators. 9. Conclusion It is anticipated that distribution networks will continue to see a proliferation of generation from new and RE resources in coming years. In some cases this will result in power flow congestion with thermally vulnerable components restricting the installed capacity and energy yield penetration of generators. Therefore a system that can be developed for the management of power flows within distribution networks, through the power output control of generators, could be of great benefit. The development stages of such a system were outlined in Sections 3 – 6. These included: (i) the identification of thermally vulnerable components through a thermal vulnerability factor assessment that related power flow sensitivity factors to component thermal limits, (ii) the targeted development of thermal models to allow latent capacity in the network assets to be unlocked through a real-time thermal rating system, (iii) the development of thermal state estimation techniques that would allow component steady-state thermal models to be populated with real-time data from meteorological monitoring equipment installed in the distribution network (iv) the use Increasing the energy yield of generation from new and renewable energy resources 59Renewable Energy60 of the real-time thermal ratings together with power flow sensitivity factors to control the power output of multiple generators. The development stages were illustrated with a case study and the energy yields resulting from the offline application of strategies were quantified. It was demonstrated for the particular case study scenario that the adoption of both a real-time thermal rating system and power flow sensitivity factor-based control strategies have the potential to unlock gains in the aggregated annual energy yield of multiple generation installations. Work is continuing in this area to realise the potential of generator output control systems for increased new and renewable energy yield penetrations. 10. References Abu-Hamdeh, N.H. (2003). Thermal properties of soil as affected by density and water content. Biosystems engineering, Vol. 86, No.1 (September 2003) 97-102 ANSI/IEEE. (1981). C57.92 Guide for loading mineral oil-immersed power transformers up to and including 100 MVA with 55°C or 65°C average winding rise, IEEE Belben, P.D. & Ziesler, C.D. (2002). Aeolian uprating: how wind farms can solve their own transmission problems, Proceedings of 1st World Wind Energy Conference and Exhibition, Berlin, Germany, July 2002 Berende, M.J.C., Slootweg, J.G. & Clemens, G.J.M.B. (2005). Incorporating Weather Statistics in Determining Overhead Line Ampacity, Proceedings of International Conference on Future Power Systems, 1-8, ISSN 90-7805-02-4, Amsterdam, Netherlands, November 2005, FPS, Amsterdam Celia, M.; Boulotas, E. & Zarba, R. (1990). A general mass-conservative numerical solution for the unsaturated flow problem. Water resources research, Vol. 26, No.7, 1483-1496 Currie, R. A. F.; Ault, G. W. & McDonald, J. R. (2006). Methodology for determination of economic connection capacity for renewable generator connections to distribution networks optimised by active power flow management. IEE Generation, Transmission, Distribution, Vol. 153, No. 4, (July 2006), 456-462, ISSN 1350-2360 Dinic, N.; Flynn, D.; Xu, L. & Kennedy, A. (2006). Increasing Wind Farm Capacity. IEE Generation, Transmission, Distribution, Vol. 153, No. 4, (July 2006), 493-498, 1350-2360 Douglass, D. A. & Edris A. A. (1996). Real-time monitoring and dynamic thermal rating of power transmission circuits. IEEE Transactions on Power Delivery, Vol. 11, No. 3, (July 1996), 1407-1418, ISSN 0885-8977 Douglass D. A.; Edris A. A. & Pritchard G. A. (1997). Field application of a dynamic thermal circuit rating method. IEEE Transactions on Power Delivery, Vol. 12, No. 2, (April 1997), 823-831, ISSN 0885-8977 Electricity Networks Association. (1971). Engineering Recommendation P15: Transformers loading guide, ENA, London Electricity Networks Association. (1986). Engineering Recommendation P27: Current rating guide for high voltage overhead lines operating in the UK distribution system, ENA, London Electricity Networks Association. (2003). Engineering Technical Report 124: Guidelines for actively managing power flows associated with the connection of a single distributed generation plant, ENA, London Electricity Networks Association. (2004). Engineering Recommendation P17: Current rating for distribution cables, ENA, London Foote, C.; Djapic, P.; Ault, G.; Mutale, J.; Burt G. & Strbac, G. (2006). United Kingdom Generic Distribution System (UKGDS) - Summary of EHV Networks, DTI Centre for Distributed Generation and Sustainable Electrical Energy, UK Fox-Penner, P.S. (2001). Easing Gridlock on the Grid - Electricity Planning and Siting Compacts. The Electricity Journal, Vol. 14, No.9 (November 2001), 11-30 Harrison, G.P. & Wallace, A.R. (2005). Optimal Power Flow Evaluation of Distribution Network Capacity for the Connection of Distributed Generation. IEE Generation, Transmission, Distribution, Vol. 152, No. 1, (January 2005), 115-122, ISSN 1350-2360 Helmer, M. (2000). Optimized size of wind power plant transformer and parallel operation, Proceedings of Wind Power for the 21st century, Kassel, Germany, September 2000 Hoffmann, S.P. & Clark, A.M. (2004). The approach to thermal uprating of transmission lines in the UK. Electra, (B2-317), CIGRE, Paris House, H.E. & Tuttle, P. D. (1959). Current Carrying Capacity of ACSR. IEEE Transactions on Power Apparatus Systems, Vol. 78, No. 3, (April 1958), 1169-1177, ISSN 0018-9510 IEC. (1991). IEC 60826: Loading and strength of overhead transmission lines, IEC, Switzerland IEC. (1994). IEC 60287: Electric cables – calculation of the current rating, IEC, Switzerland IEC. (1995). IEC TR 1597: Overhead electrical conductors – calculation methods for stranded bare conductors, IEC, Switzerland IEC. (2008). IEC 60076-7: Power transformers - Part 7: Loading guide for oil-immersed power transformers, IEC, Switzerland IEEE. (1993). IEEE 738: Standard for calculating the current-temperature relationship of bare overhead conductors, IEEE IEEE. (1994). IEEE 835, IEEE Standard Power Cable Ampacity Tables, IEEE Ippolito, L., Vaccaro, A., Villacci, D. (2004). The use of affine arithmetic for thermal state estimation of substation distribution transformers. International Journal for Computation and Mathematics in Electrical and Electronic Engineering, Vol. 23, No. 1, 237–249, ISSN 0332-1649 Jupe, S.C.E. & Taylor, P.C. (2009a). Strategies for the control of multiple distributed generation schemes, Proceedings of 20th International Conference on Electricity Distribution, Prague, Czech Republic, June 2009, (In press) Jupe, S.C.E. & Taylor, P.C. (2009b). Distributed generation output control for network power flow management. IET Renewable Power Generation, (In press) Mendez Quezada, V.H.; Abbad, J.R. & San Romain, G.T. (2006). Assessment of Energy Distribution Losses for Increasing Penetration of Distributed Generation. IEEE Transactions on Power Systems, Vol. 21, No. 2, (May 2006), 533-540, ISSN 0885-8950 Michiorri, A.; Taylor, P.C.; Jupe, S.C.E. & Berry, C (2009). An investigation into the effect of environmental conditions on power system ratings. IMechE: Journal of Power and Energy, (In press) Morgan, V.T. (1982). The thermal rating of overhead-line conductors. Electric Power Systems Research, Vol. 5, No. 2, (June 1982), 119-139, ISSN 0378-7796 Nairen, D.; Qinyun, L. & Zhaohong, F. (2004). Heat transfer in ground heat exchangers with groundwater advection. International Journal of Thermal Sciences, Vol. 43, (December 2004) 1203-1211 Increasing the energy yield of generation from new and renewable energy resources 61Renewable Energy60 of the real-time thermal ratings together with power flow sensitivity factors to control the power output of multiple generators. The development stages were illustrated with a case study and the energy yields resulting from the offline application of strategies were quantified. It was demonstrated for the particular case study scenario that the adoption of both a real-time thermal rating system and power flow sensitivity factor-based control strategies have the potential to unlock gains in the aggregated annual energy yield of multiple generation installations. Work is continuing in this area to realise the potential of generator output control systems for increased new and renewable energy yield penetrations. 10. References Abu-Hamdeh, N.H. (2003). Thermal properties of soil as affected by density and water content. Biosystems engineering, Vol. 86, No.1 (September 2003) 97-102 ANSI/IEEE. (1981). C57.92 Guide for loading mineral oil-immersed power transformers up to and including 100 MVA with 55°C or 65°C average winding rise, IEEE Belben, P.D. & Ziesler, C.D. (2002). Aeolian uprating: how wind farms can solve their own transmission problems, Proceedings of 1st World Wind Energy Conference and Exhibition, Berlin, Germany, July 2002 Berende, M.J.C., Slootweg, J.G. & Clemens, G.J.M.B. (2005). Incorporating Weather Statistics in Determining Overhead Line Ampacity, Proceedings of International Conference on Future Power Systems, 1-8, ISSN 90-7805-02-4, Amsterdam, Netherlands, November 2005, FPS, Amsterdam Celia, M.; Boulotas, E. & Zarba, R. (1990). A general mass-conservative numerical solution for the unsaturated flow problem. Water resources research, Vol. 26, No.7, 1483-1496 Currie, R. A. F.; Ault, G. W. & McDonald, J. R. (2006). Methodology for determination of economic connection capacity for renewable generator connections to distribution networks optimised by active power flow management. IEE Generation, Transmission, Distribution, Vol. 153, No. 4, (July 2006), 456-462, ISSN 1350-2360 Dinic, N.; Flynn, D.; Xu, L. & Kennedy, A. (2006). Increasing Wind Farm Capacity. IEE Generation, Transmission, Distribution, Vol. 153, No. 4, (July 2006), 493-498, 1350-2360 Douglass, D. A. & Edris A. A. (1996). Real-time monitoring and dynamic thermal rating of power transmission circuits. IEEE Transactions on Power Delivery, Vol. 11, No. 3, (July 1996), 1407-1418, ISSN 0885-8977 Douglass D. A.; Edris A. A. & Pritchard G. A. (1997). Field application of a dynamic thermal circuit rating method. IEEE Transactions on Power Delivery, Vol. 12, No. 2, (April 1997), 823-831, ISSN 0885-8977 Electricity Networks Association. (1971). Engineering Recommendation P15: Transformers loading guide, ENA, London Electricity Networks Association. (1986). Engineering Recommendation P27: Current rating guide for high voltage overhead lines operating in the UK distribution system, ENA, London Electricity Networks Association. (2003). Engineering Technical Report 124: Guidelines for actively managing power flows associated with the connection of a single distributed generation plant, ENA, London Electricity Networks Association. (2004). Engineering Recommendation P17: Current rating for distribution cables, ENA, London Foote, C.; Djapic, P.; Ault, G.; Mutale, J.; Burt G. & Strbac, G. (2006). United Kingdom Generic Distribution System (UKGDS) - Summary of EHV Networks, DTI Centre for Distributed Generation and Sustainable Electrical Energy, UK Fox-Penner, P.S. (2001). Easing Gridlock on the Grid - Electricity Planning and Siting Compacts. The Electricity Journal, Vol. 14, No.9 (November 2001), 11-30 Harrison, G.P. & Wallace, A.R. (2005). Optimal Power Flow Evaluation of Distribution Network Capacity for the Connection of Distributed Generation. IEE Generation, Transmission, Distribution, Vol. 152, No. 1, (January 2005), 115-122, ISSN 1350-2360 Helmer, M. (2000). Optimized size of wind power plant transformer and parallel operation, Proceedings of Wind Power for the 21st century, Kassel, Germany, September 2000 Hoffmann, S.P. & Clark, A.M. (2004). The approach to thermal uprating of transmission lines in the UK. Electra, (B2-317), CIGRE, Paris House, H.E. & Tuttle, P. D. (1959). Current Carrying Capacity of ACSR. IEEE Transactions on Power Apparatus Systems, Vol. 78, No. 3, (April 1958), 1169-1177, ISSN 0018-9510 IEC. (1991). IEC 60826: Loading and strength of overhead transmission lines, IEC, Switzerland IEC. (1994). IEC 60287: Electric cables – calculation of the current rating, IEC, Switzerland IEC. (1995). IEC TR 1597: Overhead electrical conductors – calculation methods for stranded bare conductors, IEC, Switzerland IEC. (2008). IEC 60076-7: Power transformers - Part 7: Loading guide for oil-immersed power transformers, IEC, Switzerland IEEE. (1993). IEEE 738: Standard for calculating the current-temperature relationship of bare overhead conductors, IEEE IEEE. (1994). IEEE 835, IEEE Standard Power Cable Ampacity Tables, IEEE Ippolito, L., Vaccaro, A., Villacci, D. (2004). The use of affine arithmetic for thermal state estimation of substation distribution transformers. International Journal for Computation and Mathematics in Electrical and Electronic Engineering, Vol. 23, No. 1, 237–249, ISSN 0332-1649 Jupe, S.C.E. & Taylor, P.C. (2009a). Strategies for the control of multiple distributed generation schemes, Proceedings of 20th International Conference on Electricity Distribution, Prague, Czech Republic, June 2009, (In press) Jupe, S.C.E. & Taylor, P.C. (2009b). Distributed generation output control for network power flow management. IET Renewable Power Generation, (In press) Mendez Quezada, V.H.; Abbad, J.R. & San Romain, G.T. (2006). Assessment of Energy Distribution Losses for Increasing Penetration of Distributed Generation. IEEE Transactions on Power Systems, Vol. 21, No. 2, (May 2006), 533-540, ISSN 0885-8950 Michiorri, A.; Taylor, P.C.; Jupe, S.C.E. & Berry, C (2009). An investigation into the effect of environmental conditions on power system ratings. IMechE: Journal of Power and Energy, (In press) Morgan, V.T. (1982). The thermal rating of overhead-line conductors. Electric Power Systems Research, Vol. 5, No. 2, (June 1982), 119-139, ISSN 0378-7796 Nairen, D.; Qinyun, L. & Zhaohong, F. (2004). Heat transfer in ground heat exchangers with groundwater advection. International Journal of Thermal Sciences, Vol. 43, (December 2004) 1203-1211 Increasing the energy yield of generation from new and renewable energy resources 61Renewable Energy62 Neher, J.H. & McGrath, M.H. (1957). The calculation of the temperature rise and load capability of cable systems. AIEE Transactions, Vol. 76, No. 3, (April 1957), 752-772, ISSN 0018-9510 Neumann, A.; Taylor, P.; Jupe, S.; Michiorri, A.; Goode, A.; Curry, D. & Roberts, D. (2008). Dynamic thermal rating and active control for improved distribution network utilisation, Proceedings of PowerGrid Europe, Milan, Italy, June 2008, PennWell, Milan Nuijten, J.M.A. & Geschiere, A. (2005). Future network planning and grid control, Proceedings of International Conference on Future Power Systems, pp. 7-13, ISSN 90-78205-02-4, Amsterdam, Netherlands, November 2005, FPS, Amsterdam Piccolo, A.; Vaccaro, A. & Villacci, D. (2004). Thermal rating assessment of overhead lines by Affine Arithmetic. Electric Power Systems Research, Vol. 71, No. 3, (November 2004), 275-283, ISSN 0378-7796 Roberts, D.A. (2004). Network Management Systems for Active Distribution Networks – A Feasibility Study, Scottish Power Plc, Glasgow Rodriguez-Iturbe, I.; Isham, I.; Cox D.R. et al. (2006). Space-time modelling of soil moisture: stochastic rainfall forcing with heterogeneous vegetation. Water resources research, Vol. 42, (March 2006) Shepard D. (1968). A two-dimensional interpolation function for irregularly-spaced data, Proceedings of 23rd ACM national conference, 517-524, August 1968, ACM, New York Soto, F., Latorre, J. & Wagensberg, M. (1998). Increasing the capacity of overhead lines in the 400 kV Spanish transmission network: real time thermal ratings. Electra, (22-211), CIGRE, Paris Stephen, R. (2004). Description and evaluation of options relating to uprating of overhead transmission lines. Electra, (B2-201), CIGRE, Paris Susa, D.; Lehtonen, M. & Nordman, H. (2005). Dynamic thermal modelling of power transformers. IEEE Transactions on Power Delivery, Vol. 20, No. 1, (January 2005), 197-204, ISSN 0885-8977 Van Genuchten M. (1980). A closed for equation for predicting the hydraulic conductivity of unsaturated soils. Soil science society of America Journal, Vol. 44, (May 1980), 892-898 Villacci, D. & Vaccaro, A. (2007). Transient tolerance analysis of power cables thermal dynamic by interval mathematic. Electric Power Systems Research, Vol. 77, No. 3-4, (March 2007), 308–314, ISSN 0378-7796 Vovos, P. N.; Harrison, G.P.; Wallace, A.R. & Bialek, J. W. (2005). Optimal Power Flow as a Tool for Fault Level-Constrained Network Capacity Analysis. IEEE Transactions on Power Systems, Vol. 20, No. 2, (May 2005), 734-741, ISSN 0885-8950 WG 22.12. (1992). The Thermal Behaviour of Overhead Conductors – Section 1. Electra, Vol. 144, No. 3, CIGRE, Paris Wood, A.J. & Wollenberg, B.F. (1996). Power Generation, Operation and Control, Wiley, ISBN 0-471-58699-4, New York, USA Embedded Energy Storage Systems in the Power Grid for Renewable Energy Sources Integration 63Embedded Energy Storage Systems in the Power Grid for Renewable Energy Sources IntegrationSérgio Faias, Jorge Sousa and Rui CastroX Embedded Energy Storage Systems in the Power Grid for Renewable Energy Sources Integration Sérgio Faias*, Jorge Sousa* and Rui Castro** * ISEL –Instituto Superior de Engenharia de Lisboa ** Cie3 / IST – Technical University of Lisbon Portugal 1. Introduction The increasing integration of renewable energy sources in the power system allows a reduction of the fossil fuels consumption and on their consequent environmental impacts, such as the climate change and air pollution. Moreover, a high share of renewable power allows other significant advantages for the countries promoting these technologies, such as the improvement of the security of supply and the decrease of the energy dependency from other countries. Nevertheless, as the most renewable technologies ultimately derive energy from unpredictable natural sources or sources that vary in their availability over different timescales, the integration of larger amounts of renewable energy raises some issues related to the power systems operation (IEA, 2005). In this regard, the power demand and generation balancing becomes one of most important issues that result from the increasing integration of renewable sources. This issue will emerge especially in periods with high renewable generation and low demand and implies an action from the power system operator in order to maintain the system stability. Apparently, there are three different solutions to face the power demand and generation balancing; the first one is the curtailment of the renewable generation, resulting in a waste of this power source, the second one is to export power to an interconnected grid (Hammons, 2006), not disregarding the possibility of that grid being also submitted to the same power balance issue, and finally the third solution, that avoids the disadvantages of the preceding ones, is to store the excess of renewable power for use when it is needed, keeping the stability of the power grid (Barton & Infield, 2004 and Faias et al., 2007). The concept of energy storage systems embedded in the power grid, which will be presented in the current chapter, allows decoupling partially the demand from the supply, facilitating the penetration of renewable sources and improving the flexibility and efficiency of the power grid. This concept will be applied to the Portuguese power system case, considering the predictions for the load and for the generation installed capacity, including renewable, for the year 2011. 5Renewable Energy62 Neher, J.H. & McGrath, M.H. (1957). The calculation of the temperature rise and load capability of cable systems. AIEE Transactions, Vol. 76, No. 3, (April 1957), 752-772, ISSN 0018-9510 Neumann, A.; Taylor, P.; Jupe, S.; Michiorri, A.; Goode, A.; Curry, D. & Roberts, D. (2008). Dynamic thermal rating and active control for improved distribution network utilisation, Proceedings of PowerGrid Europe, Milan, Italy, June 2008, PennWell, Milan Nuijten, J.M.A. & Geschiere, A. (2005). Future network planning and grid control, Proceedings of International Conference on Future Power Systems, pp. 7-13, ISSN 90-78205-02-4, Amsterdam, Netherlands, November 2005, FPS, Amsterdam Piccolo, A.; Vaccaro, A. & Villacci, D. (2004). Thermal rating assessment of overhead lines by Affine Arithmetic. Electric Power Systems Research, Vol. 71, No. 3, (November 2004), 275-283, ISSN 0378-7796 Roberts, D.A. (2004). Network Management Systems for Active Distribution Networks – A Feasibility Study, Scottish Power Plc, Glasgow Rodriguez-Iturbe, I.; Isham, I.; Cox D.R. et al. (2006). Space-time modelling of soil moisture: stochastic rainfall forcing with heterogeneous vegetation. Water resources research, Vol. 42, (March 2006) Shepard D. (1968). A two-dimensional interpolation function for irregularly-spaced data, Proceedings of 23rd ACM national conference, 517-524, August 1968, ACM, New York Soto, F., Latorre, J. & Wagensberg, M. (1998). Increasing the capacity of overhead lines in the 400 kV Spanish transmission network: real time thermal ratings. Electra, (22-211), CIGRE, Paris Stephen, R. (2004). Description and evaluation of options relating to uprating of overhead transmission lines. Electra, (B2-201), CIGRE, Paris Susa, D.; Lehtonen, M. & Nordman, H. (2005). Dynamic thermal modelling of power transformers. IEEE Transactions on Power Delivery, Vol. 20, No. 1, (January 2005), 197-204, ISSN 0885-8977 Van Genuchten M. (1980). A closed for equation for predicting the hydraulic conductivity of unsaturated soils. Soil science society of America Journal, Vol. 44, (May 1980), 892-898 Villacci, D. & Vaccaro, A. (2007). Transient tolerance analysis of power cables thermal dynamic by interval mathematic. Electric Power Systems Research, Vol. 77, No. 3-4, (March 2007), 308–314, ISSN 0378-7796 Vovos, P. N.; Harrison, G.P.; Wallace, A.R. & Bialek, J. W. (2005). Optimal Power Flow as a Tool for Fault Level-Constrained Network Capacity Analysis. IEEE Transactions on Power Systems, Vol. 20, No. 2, (May 2005), 734-741, ISSN 0885-8950 WG 22.12. (1992). The Thermal Behaviour of Overhead Conductors – Section 1. Electra, Vol. 144, No. 3, CIGRE, Paris Wood, A.J. & Wollenberg, B.F. (1996). Power Generation, Operation and Control, Wiley, ISBN 0-471-58699-4, New York, USA Embedded Energy Storage Systems in the Power Grid for Renewable Energy Sources Integration 63Embedded Energy Storage Systems in the Power Grid for Renewable Energy Sources IntegrationSérgio Faias, Jorge Sousa and Rui CastroX Embedded Energy Storage Systems in the Power Grid for Renewable Energy Sources Integration Sérgio Faias*, Jorge Sousa* and Rui Castro** * ISEL –Instituto Superior de Engenharia de Lisboa ** Cie3 / IST – Technical University of Lisbon Portugal 1. Introduction The increasing integration of renewable energy sources in the power system allows a reduction of the fossil fuels consumption and on their consequent environmental impacts, such as the climate change and air pollution. Moreover, a high share of renewable power allows other significant advantages for the countries promoting these technologies, such as the improvement of the security of supply and the decrease of the energy dependency from other countries. Nevertheless, as the most renewable technologies ultimately derive energy from unpredictable natural sources or sources that vary in their availability over different timescales, the integration of larger amounts of renewable energy raises some issues related to the power systems operation (IEA, 2005). In this regard, the power demand and generation balancing becomes one of most important issues that result from the increasing integration of renewable sources. This issue will emerge especially in periods with high renewable generation and low demand and implies an action from the power system operator in order to maintain the system stability. Apparently, there are three different solutions to face the power demand and generation balancing; the first one is the curtailment of the renewable generation, resulting in a waste of this power source, the second one is to export power to an interconnected grid (Hammons, 2006), not disregarding the possibility of that grid being also submitted to the same power balance issue, and finally the third solution, that avoids the disadvantages of the preceding ones, is to store the excess of renewable power for use when it is needed, keeping the stability of the power grid (Barton & Infield, 2004 and Faias et al., 2007). The concept of energy storage systems embedded in the power grid, which will be presented in the current chapter, allows decoupling partially the demand from the supply, facilitating the penetration of renewable sources and improving the flexibility and efficiency of the power grid. This concept will be applied to the Portuguese power system case, considering the predictions for the load and for the generation installed capacity, including renewable, for the year 2011. 5Renewable Energy64 The methodology used along this chapter begins with the description of the power unbalance between generation and demand issue and the solutions for its mitigation. After that, the available energy storage technologies are presented as well as the energy storage system design process. Following, a brief discussion on the different perspectives for the location of embedded energy storage system in the power grid is presented. Finally, the results for Portuguese power system case study are presented, analysed and discussed. 2. Balancing Power Demand and Generation Traditionally the power system operation is based on the principle that at each moment in time the power demanded by the load is generated, at that moment, by the set of power plants in the system. For that reason, if either the demand or the generation experience a sudden and sharp increase or decrease, power unbalances would occur resulting in power system instability. As, in general, the power demand is not controlled by the power system operator, its action is focused on the power generation dispatching. Power generation dispatch is also known as the unit commitment, and it consists of defining which power units should be operating at a specific moment and the power generated by each unit (dispatch) in order to satisfy the power demand. Power demand varies along time and according to different drivers, such as seasonality, weather condition, geographical location and the society economic activity. Extreme weather temperatures usually induce a higher power demand. On the other hand, there is a reduction on the power demand during the night and during the weekends, as a result of lower industrial activity and people’s specific daily routine. Figure 1 presents different time cycles variations imposed by each one of the causes referred above. Fig. 1. Power demand variation along different time cycles (REN, 2008) In the past, from the power system operator perspective, the uncertainly was mainly on the demand-side. However, the increasing integration of non-dispatchable renewable power sources results in an uncertainly that is also endogenous to the generation-side. Renewable energy, which is dependent on natural resources like wind, rain and sun, present a much variable availability and can not be easily committed and dispatched by the power system operator. This variability of renewable power is apparent in Figure 2, where an example of the aggregated wind power generation along the period of one week is presented. Fig. 2. Aggregated wind power generation along the period of one week (REN, 2008) Therefore, the system operator usually considers the renewable generation as a negative power demand, satisfying the remaining demand through the unit commitment of the conventional (thermal and large hydro) power plants (Ortega-Vazquez & Kirschen, 2009). 2.1 Power Unbalance Besides the power system operation constraints induced by the renewable power sources availability, there is an additional issue that should be taken into account. This issue is the power unbalance between power generation and demand, resulting from the coincidence of the higher availability of renewable sources (wind power, for instance) with the periods of lower demand. This fact is confirmed in Figure 3, where the equivalent average day of an annual aggregated wind power generation and the average day of an annual aggregated power demand are compared. Fig. 3. Comparison between the equivalent average day of an annual aggregated wind power generation and the equivalent average day of an annual aggregated power demand (REN, 2008) The coincidence of the wind peak power with the demand off-peak power may origin some moments where the non-dispatchable generation is greater than the demand. The resulting power unbalance can be characterized by: Embedded Energy Storage Systems in the Power Grid for Renewable Energy Sources Integration 65Renewable Energy64 The methodology used along this chapter begins with the description of the power unbalance between generation and demand issue and the solutions for its mitigation. After that, the available energy storage technologies are presented as well as the energy storage system design process. Following, a brief discussion on the different perspectives for the location of embedded energy storage system in the power grid is presented. Finally, the results for Portuguese power system case study are presented, analysed and discussed. 2. Balancing Power Demand and Generation Traditionally the power system operation is based on the principle that at each moment in time the power demanded by the load is generated, at that moment, by the set of power plants in the system. For that reason, if either the demand or the generation experience a sudden and sharp increase or decrease, power unbalances would occur resulting in power system instability. As, in general, the power demand is not controlled by the power system operator, its action is focused on the power generation dispatching. Power generation dispatch is also known as the unit commitment, and it consists of defining which power units should be operating at a specific moment and the power generated by each unit (dispatch) in order to satisfy the power demand. Power demand varies along time and according to different drivers, such as seasonality, weather condition, geographical location and the society economic activity. Extreme weather temperatures usually induce a higher power demand. On the other hand, there is a reduction on the power demand during the night and during the weekends, as a result of lower industrial activity and people’s specific daily routine. Figure 1 presents different time cycles variations imposed by each one of the causes referred above. Fig. 1. Power demand variation along different time cycles (REN, 2008) In the past, from the power system operator perspective, the uncertainly was mainly on the demand-side. However, the increasing integration of non-dispatchable renewable power sources results in an uncertainly that is also endogenous to the generation-side. Renewable energy, which is dependent on natural resources like wind, rain and sun, present a much variable availability and can not be easily committed and dispatched by the power system operator. This variability of renewable power is apparent in Figure 2, where an example of the aggregated wind power generation along the period of one week is presented. Fig. 2. Aggregated wind power generation along the period of one week (REN, 2008) Therefore, the system operator usually considers the renewable generation as a negative power demand, satisfying the remaining demand through the unit commitment of the conventional (thermal and large hydro) power plants (Ortega-Vazquez & Kirschen, 2009). 2.1 Power Unbalance Besides the power system operation constraints induced by the renewable power sources availability, there is an additional issue that should be taken into account. This issue is the power unbalance between power generation and demand, resulting from the coincidence of the higher availability of renewable sources (wind power, for instance) with the periods of lower demand. This fact is confirmed in Figure 3, where the equivalent average day of an annual aggregated wind power generation and the average day of an annual aggregated power demand are compared. Fig. 3. Comparison between the equivalent average day of an annual aggregated wind power generation and the equivalent average day of an annual aggregated power demand (REN, 2008) The coincidence of the wind peak power with the demand off-peak power may origin some moments where the non-dispatchable generation is greater than the demand. The resulting power unbalance can be characterized by: Embedded Energy Storage Systems in the Power Grid for Renewable Energy Sources Integration 65Renewable Energy66      tPtPtP DemGenUnb  (1) where  tPUnb is the power unbalance,  tPGen is the power generated and  tPDem is the power demand, at time t. Positive power unbalances correspond to an excess of generation, while negative power unbalances correspond to a generation shortage. Positive power unbalances can occur in power systems with high penetration of non-dispatchable renewable power sources, like wind power. High hydro power availability and the minimal spinning reserve of the thermal power units also contribute to the occurrence of positive power unbalances. Negative power unbalances only occur in moments when the available power capacity is not enough to cover all the power demands needs. 2.2 Solutions for Power Unbalance Mitigation In practice no power unbalances should occur in a power system, otherwise, it would not be possible to the power system operator to keep the system in perfect operation and maintaining the standard power quality levels. In moments when the power system tends to be unbalanced, namely driven by an increasing renewable power generation, the power system operator must act in order to mitigate that power unbalance and its consequences. Following, it is presented a sort of solutions that can be adopted, individually or complementarily, to mitigate the power unbalance issue, such as curtailment of renewable generation, interconnections with other power systems and energy storage (a) Curtailment of renewable generation The curtailment of renewable generation is one of the possible solutions to avoid unbalances between power generation and power demand. Such solution consists of an order emitted by the power system operator to the renewable power producers to cut partially or totally their generation. Renewable generation curtailment usually implies the waste of an environmental friendly natural resource and an increase on the fossil fuels consumption. As so, this power unbalance mitigation solution should only be considered in case of extreme contingencies. (b) Interconnections with other power systems Strong interconnections between different power systems are today an important advantage in terms of energy management and compensation of local power unbalances (Hammons, 2006). The recent increasing integration of renewable power sources has been one of the drivers to reinforce the interconnection capacity of the power grids, enabling each country or region to best exploit the endogenous renewable resources creating together a diversified generation mix. In spite of the interconnection between different power systems being one of the best solutions to mitigate the power unbalance, there are some constraints to its application, namely the one related to the potential of existing simultaneous power unbalances in the interconnected regions. (c) Energy storage Energy storage offers additional benefits in utility settings because it can decouple demand from supply, thereby mitigating the unbalances on the power system and allowing increased asset utilization, facilitating the penetration of renewable sources, and improving the flexibility, reliability, and efficiency of the electrical network (Schoenung, 1996). The option for energy storage solution involves an investment on an energy storage system, but, on the other hand, it avoids the disadvantage of the renewable power curtailment and the constraints related to exporting power through the grid interconnections. 3. Energy Storage Technologies Today, there are several high performance storage technologies available or at an advanced state of development, demanded by a new range of the energy storage applications. The singularities of each storage technology, dependent on their operation fundamentals, turn it unique and difficult to compare them. A typical method used for the storage technology comparison, based on the power and energy capacities of commercialized devices, is presented as example in Figure 4. This power and energy range comparison of the technologies allows the identification of the devices that are best suited for a specific application. Fig. 4. Power and energy capacity comparison for different energy storage technologies (ESA, 2009) In order to best distinguish the energy storage technologies applications and considering their placement on the ordinates axe of the figure above, storage technologies are here classified into two different categories, based on their discharge time. These categories are: short-term discharge energy storage devices and long-term discharge energy storage devices. Short-term discharge energy storage devices present a very fast response to the power system needs. However, they just can supply their rated power for short periods, which Embedded Energy Storage Systems in the Power Grid for Renewable Energy Sources Integration 67Renewable Energy66      tPtPtP DemGenUnb  (1) where  tPUnb is the power unbalance,  tPGen is the power generated and  tPDem is the power demand, at time t. Positive power unbalances correspond to an excess of generation, while negative power unbalances correspond to a generation shortage. Positive power unbalances can occur in power systems with high penetration of non-dispatchable renewable power sources, like wind power. High hydro power availability and the minimal spinning reserve of the thermal power units also contribute to the occurrence of positive power unbalances. Negative power unbalances only occur in moments when the available power capacity is not enough to cover all the power demands needs. 2.2 Solutions for Power Unbalance Mitigation In practice no power unbalances should occur in a power system, otherwise, it would not be possible to the power system operator to keep the system in perfect operation and maintaining the standard power quality levels. In moments when the power system tends to be unbalanced, namely driven by an increasing renewable power generation, the power system operator must act in order to mitigate that power unbalance and its consequences. Following, it is presented a sort of solutions that can be adopted, individually or complementarily, to mitigate the power unbalance issue, such as curtailment of renewable generation, interconnections with other power systems and energy storage (a) Curtailment of renewable generation The curtailment of renewable generation is one of the possible solutions to avoid unbalances between power generation and power demand. Such solution consists of an order emitted by the power system operator to the renewable power producers to cut partially or totally their generation. Renewable generation curtailment usually implies the waste of an environmental friendly natural resource and an increase on the fossil fuels consumption. As so, this power unbalance mitigation solution should only be considered in case of extreme contingencies. (b) Interconnections with other power systems Strong interconnections between different power systems are today an important advantage in terms of energy management and compensation of local power unbalances (Hammons, 2006). The recent increasing integration of renewable power sources has been one of the drivers to reinforce the interconnection capacity of the power grids, enabling each country or region to best exploit the endogenous renewable resources creating together a diversified generation mix. In spite of the interconnection between different power systems being one of the best solutions to mitigate the power unbalance, there are some constraints to its application, namely the one related to the potential of existing simultaneous power unbalances in the interconnected regions. (c) Energy storage Energy storage offers additional benefits in utility settings because it can decouple demand from supply, thereby mitigating the unbalances on the power system and allowing increased asset utilization, facilitating the penetration of renewable sources, and improving the flexibility, reliability, and efficiency of the electrical network (Schoenung, 1996). The option for energy storage solution involves an investment on an energy storage system, but, on the other hand, it avoids the disadvantage of the renewable power curtailment and the constraints related to exporting power through the grid interconnections. 3. Energy Storage Technologies Today, there are several high performance storage technologies available or at an advanced state of development, demanded by a new range of the energy storage applications. The singularities of each storage technology, dependent on their operation fundamentals, turn it unique and difficult to compare them. A typical method used for the storage technology comparison, based on the power and energy capacities of commercialized devices, is presented as example in Figure 4. This power and energy range comparison of the technologies allows the identification of the devices that are best suited for a specific application. Fig. 4. Power and energy capacity comparison for different energy storage technologies (ESA, 2009) In order to best distinguish the energy storage technologies applications and considering their placement on the ordinates axe of the figure above, storage technologies are here classified into two different categories, based on their discharge time. These categories are: short-term discharge energy storage devices and long-term discharge energy storage devices. Short-term discharge energy storage devices present a very fast response to the power system needs. However, they just can supply their rated power for short periods, which Embedded Energy Storage Systems in the Power Grid for Renewable Energy Sources Integration 67Renewable Energy68 vary from milliseconds to few minutes. The short-term discharge energy storage devices are usually applied to improve power quality, to cover load during start-up and synchronization of backup generators and to compensate transient response of renewable power sources (Tande, 2003). Long-term discharge energy storage devices are able to supply power from some seconds to many hours. Their response to the power system needs is usually slower than the short-term discharge energy storage devices, and much dependent on the technology. Long-term discharge energy storage devices are usually applied on the energy management, renewable energy sources integration and power grid congestion management (Price et al., 1999). 3.1 Short-term discharge energy storage devices Short-term discharge energy storage devices should be used to aid power systems during the transient period after a system disturbance, such as line switching, load changes and fault clearance. Their application prevents collapse of power systems due to loss of synchronism or voltage instability, improving its reliability and quality. Short-term discharge energy storage devices use is getting common in power systems with important renewable energy penetration (like wind, for instance) and weak interconnections or in islands, avoiding temporary faults and contributing to the provision of important system services such as momentary reserves and short-circuit capacity (Hamsic et al., 2007). The main short-term discharge energy storage devices and their operation are presented below. (a) Flywheels Flywheels store kinetic energy in a rotating mass. Such equipments have typically been used as short-term energy storage devices for propulsion applications such as powering train engines and road vehicles, and in centrifuges. In these applications, the flywheel smoothes the power load during deceleration by dynamic braking action and then provides a boost during acceleration (Lazarewicz and Rojas, 2004). Figure 5 presents the operating diagram of a flywheel energy storage system. Fig. 5. Flywheel energy storage device operation diagram (b) Supercapacitors Supercapacitors are the latest innovative devices in the field of electrical energy storage. In comparison with a battery or a traditional capacitor, the supercapacitor allows a much powerful power and energy density (Zhai et al., 2006). Supercapacitors are electrochemical double layer capacitors that store energy as electric charge between two plates, metal or conductive, separated by a dielectric, when a voltage differential is applied across the plates (Rufer et al., 2004). As like battery systems, capacitors work in direct current. This fact imposes the use of electronic power systems, as presented in Figure 6. Fig. 6. Supercapacitor energy storage device operation diagram (b) Magnetic Superconducting Superconducting magnetic energy storage devices store energy in the form of a magnetic field, through a direct current flowing in a superconducting coil. The alternate current from a power bus is converted to direct current and injected in the coil. When necessary, the stored energy can be released, through a direct current that is converted to alternate current and injected in the power bus (Hsu and Lee, 1992). The interface between the power bus and the superconducting coil uses power electronic converters (Nomura et al., 2006). The Superconducting Magnetic Energy Storage (SMES) device operation diagram is presented in Figure 7. Fig. 7. SMES device operation diagram The conductor for carrying the direct current operates at cryogenic temperatures where it behaves as a superconductor and thus has virtually no resistive losses as it produces the magnetic field. Consequently, the energy can be stored in a persistent mode, until required. The most important advantage of SMES device is that the time delay during charge and discharge is quite short. Power is available almost instantaneously and very high power output can be provided for a brief period of time (Mito et al., 2004). Embedded Energy Storage Systems in the Power Grid for Renewable Energy Sources Integration 69Renewable Energy68 vary from milliseconds to few minutes. The short-term discharge energy storage devices are usually applied to improve power quality, to cover load during start-up and synchronization of backup generators and to compensate transient response of renewable power sources (Tande, 2003). Long-term discharge energy storage devices are able to supply power from some seconds to many hours. Their response to the power system needs is usually slower than the short-term discharge energy storage devices, and much dependent on the technology. Long-term discharge energy storage devices are usually applied on the energy management, renewable energy sources integration and power grid congestion management (Price et al., 1999). 3.1 Short-term discharge energy storage devices Short-term discharge energy storage devices should be used to aid power systems during the transient period after a system disturbance, such as line switching, load changes and fault clearance. Their application prevents collapse of power systems due to loss of synchronism or voltage instability, improving its reliability and quality. Short-term discharge energy storage devices use is getting common in power systems with important renewable energy penetration (like wind, for instance) and weak interconnections or in islands, avoiding temporary faults and contributing to the provision of important system services such as momentary reserves and short-circuit capacity (Hamsic et al., 2007). The main short-term discharge energy storage devices and their operation are presented below. (a) Flywheels Flywheels store kinetic energy in a rotating mass. Such equipments have typically been used as short-term energy storage devices for propulsion applications such as powering train engines and road vehicles, and in centrifuges. In these applications, the flywheel smoothes the power load during deceleration by dynamic braking action and then provides a boost during acceleration (Lazarewicz and Rojas, 2004). Figure 5 presents the operating diagram of a flywheel energy storage system. Fig. 5. Flywheel energy storage device operation diagram (b) Supercapacitors Supercapacitors are the latest innovative devices in the field of electrical energy storage. In comparison with a battery or a traditional capacitor, the supercapacitor allows a much powerful power and energy density (Zhai et al., 2006). Supercapacitors are electrochemical double layer capacitors that store energy as electric charge between two plates, metal or conductive, separated by a dielectric, when a voltage differential is applied across the plates (Rufer et al., 2004). As like battery systems, capacitors work in direct current. This fact imposes the use of electronic power systems, as presented in Figure 6. Fig. 6. Supercapacitor energy storage device operation diagram (b) Magnetic Superconducting Superconducting magnetic energy storage devices store energy in the form of a magnetic field, through a direct current flowing in a superconducting coil. The alternate current from a power bus is converted to direct current and injected in the coil. When necessary, the stored energy can be released, through a direct current that is converted to alternate current and injected in the power bus (Hsu and Lee, 1992). The interface between the power bus and the superconducting coil uses power electronic converters (Nomura et al., 2006). The Superconducting Magnetic Energy Storage (SMES) device operation diagram is presented in Figure 7. Fig. 7. SMES device operation diagram The conductor for carrying the direct current operates at cryogenic temperatures where it behaves as a superconductor and thus has virtually no resistive losses as it produces the magnetic field. Consequently, the energy can be stored in a persistent mode, until required. The most important advantage of SMES device is that the time delay during charge and discharge is quite short. Power is available almost instantaneously and very high power output can be provided for a brief period of time (Mito et al., 2004). Embedded Energy Storage Systems in the Power Grid for Renewable Energy Sources Integration 69Renewable Energy70 3.2 Long-term discharge energy storage devices The so-called long-term discharge energy storage devices have the capability to supply or absorb electrical energy during hours. Sort of different long-term discharge energy storage technologies are already available today and their use is expected to rise in the next years because of the increasing integration of non-dispatchable renewable energy generation in the power systems (IEA, 2005). A brief description of the main long-term discharge energy storage technologies is presented below. (a) Pumping Hydro In pumping hydro energy storage, a body of water at a relatively high elevation represents a potential or stored energy. When generation is needed, the water in the upper reservoir is lead through a pipe downhill into a hydroelectric generator and stored in the lower reservoir. To recharge the storage system, the water is pumped back up to the upper reservoir and the power plant acts like a load as far as the power system is concerned. Pumping hydro energy storage system is constituted by two water reservoirs, an electric machine (motor/generator) and a reversible hydro pump-turbine unit. The system can be started-up in few minutes and its autonomy depends on the volume of stored water. There are three possible configurations for the pumping hydro systems. The first one, the pure pumping hydro, corresponds to a power plant that is specifically set-up for storage, where the only turbinated/pumped water is the one stored in the upper and lower reservoirs. The second configuration corresponds to a reservoir hydro power plant, integrated in a river course, equipped with a lower reservoir and a reversible pump-turbine unit. The third configuration corresponds to a cascade of hydro power plants, where some reservoirs act simultaneously like upper and lower reservoir for the different power plants. In second and third configurations, the most common, the power plant operation is more complex because of the coordination of the different power plants and the reservoir inflows resultant from the river. The operation of a pumping hydro system is presented in Figure 8. Fig. 8. Pumping hydro system operation diagram Pumping hydro energy storage system operation is constrained by the weather conditions, reducing its storage capacity in periods extremely wet or dry. The main restrictions to pumping hydro energy storage implementation are related with geographical constraints. (b) Batteries Batteries store energy in electrochemical form creating electrically charged ions. When the battery charges, a direct current is converted in chemical potential energy, when discharges, the chemical energy is converted back into a flow of electrons in direct current form (Hunt, 1998). The connection of the system to the grid, as presented in Figure 9, implies the use of power electronic converters in order to rectify the alternate current during the battery charge periods and to invert the direct current during the battery discharge periods. Fig. 9. Battery device operation diagram Batteries are the most popular energy storage devices. However, the term battery comprises a sort of several technologies applying different operation principals and materials. As so, the distinction between two important battery concepts, electrochemical and redox-flow, is hereby emphasized. Electrochemical Electrochemical batteries use electrodes both as part of the electron transfer process and store the products or reactants via electrode solid-state reactions (Price et al., 1999). There are a number of battery technologies under consideration for energy storage, the main being:  Lead acid  Nickel cadmium  Nickel metal-hydride  Sodium sulphur  Lithium ion Redox-Flow Redox-flow batteries are storage devices that convert electrical energy into chemical potential energy by charging two liquid electrolyte solutions and subsequently releasing the stored energy during discharge (Ponce de León et al., 2006). The name redox-flow battery is based on the redox reaction between the two electrolytes in the system. These reactions include all chemical processes in which atoms have their oxidation number changed. In a redox flow cell the two electrolytes are separated by a semi-permeable membrane. This membrane allows ion flow, but prevents mixing of the liquids. Electrical contact is made through inert conductors in the liquids. As the ions flow across the membrane, an electrical current is induced in the conductors (EPRI, 2007). Embedded Energy Storage Systems in the Power Grid for Renewable Energy Sources Integration 71Renewable Energy70 3.2 Long-term discharge energy storage devices The so-called long-term discharge energy storage devices have the capability to supply or absorb electrical energy during hours. Sort of different long-term discharge energy storage technologies are already available today and their use is expected to rise in the next years because of the increasing integration of non-dispatchable renewable energy generation in the power systems (IEA, 2005). A brief description of the main long-term discharge energy storage technologies is presented below. (a) Pumping Hydro In pumping hydro energy storage, a body of water at a relatively high elevation represents a potential or stored energy. When generation is needed, the water in the upper reservoir is lead through a pipe downhill into a hydroelectric generator and stored in the lower reservoir. To recharge the storage system, the water is pumped back up to the upper reservoir and the power plant acts like a load as far as the power system is concerned. Pumping hydro energy storage system is constituted by two water reservoirs, an electric machine (motor/generator) and a reversible hydro pump-turbine unit. The system can be started-up in few minutes and its autonomy depends on the volume of stored water. There are three possible configurations for the pumping hydro systems. The first one, the pure pumping hydro, corresponds to a power plant that is specifically set-up for storage, where the only turbinated/pumped water is the one stored in the upper and lower reservoirs. The second configuration corresponds to a reservoir hydro power plant, integrated in a river course, equipped with a lower reservoir and a reversible pump-turbine unit. The third configuration corresponds to a cascade of hydro power plants, where some reservoirs act simultaneously like upper and lower reservoir for the different power plants. In second and third configurations, the most common, the power plant operation is more complex because of the coordination of the different power plants and the reservoir inflows resultant from the river. The operation of a pumping hydro system is presented in Figure 8. Fig. 8. Pumping hydro system operation diagram Pumping hydro energy storage system operation is constrained by the weather conditions, reducing its storage capacity in periods extremely wet or dry. The main restrictions to pumping hydro energy storage implementation are related with geographical constraints. (b) Batteries Batteries store energy in electrochemical form creating electrically charged ions. When the battery charges, a direct current is converted in chemical potential energy, when discharges, the chemical energy is converted back into a flow of electrons in direct current form (Hunt, 1998). The connection of the system to the grid, as presented in Figure 9, implies the use of power electronic converters in order to rectify the alternate current during the battery charge periods and to invert the direct current during the battery discharge periods. Fig. 9. Battery device operation diagram Batteries are the most popular energy storage devices. However, the term battery comprises a sort of several technologies applying different operation principals and materials. As so, the distinction between two important battery concepts, electrochemical and redox-flow, is hereby emphasized. Electrochemical Electrochemical batteries use electrodes both as part of the electron transfer process and store the products or reactants via electrode solid-state reactions (Price et al., 1999). There are a number of battery technologies under consideration for energy storage, the main being:  Lead acid  Nickel cadmium  Nickel metal-hydride  Sodium sulphur  Lithium ion Redox-Flow Redox-flow batteries are storage devices that convert electrical energy into chemical potential energy by charging two liquid electrolyte solutions and subsequently releasing the stored energy during discharge (Ponce de León et al., 2006). The name redox-flow battery is based on the redox reaction between the two electrolytes in the system. These reactions include all chemical processes in which atoms have their oxidation number changed. In a redox flow cell the two electrolytes are separated by a semi-permeable membrane. This membrane allows ion flow, but prevents mixing of the liquids. Electrical contact is made through inert conductors in the liquids. As the ions flow across the membrane, an electrical current is induced in the conductors (EPRI, 2007). Embedded Energy Storage Systems in the Power Grid for Renewable Energy Sources Integration 71Renewable Energy72 Over the past few years three types of redox-flow batteries had been developed up to the stage of demonstration and commercialization. These types are vanadium redox batteries (VRB), the polysulphide bromide batteries (PSB) and the zinc bromine (ZnBr). (c) Compressed air Compressed air energy storage is a device based on as gas turbine where the compression and the combustion processes are divided. During charging, the compressor is coupled to the electrical machine, working as a motor, compressing the air. After the compression, the air is stored into a sealed underground cavern. Discharging the device consists in generating power through the coupling of the gas turbine with the electrical machine, working as generator, and supplying the stored compressed air to the combustion process (Lerch, 2007). A compressed air energy storage system operation diagram is presented in Figure 9. Fig. 9. Compressed air energy storage system operation diagram Three air reservoir types are generally considered: naturally occurring aquifers (such as those used for natural gas storage), solution-mined salt caverns, and mechanically formed reservoirs in rock formations. Main compressed air energy storage system implementation constraints are related with reservoirs achievement (Schoenung, 1996) (d) Hydrogen fuel cell A fuel cell is an energy conversion device that is closely related to a battery. Both are electrochemical devices for the conversion of chemical to electrical energy. In a battery the chemical energy is stored internally, whereas in a fuel cell the chemical energy (fuel and oxidant) is supplied externally and can be continuously replenished (Hoogers, 2003). The overall reaction in a fuel cell is the spontaneous reaction of hydrogen and oxygen to produce electricity and water. During the operation of a fuel cell, hydrogen is ionized into protons and electrons at the anode, the hydrogen ions are transported through the electrolyte to the cathode by an external circuit (load). At the cathode, oxygen combines with the hydrogen ions and electrons to produce water. The hydrogen fuel cell system can be reversible, allowing electric power consumption for the production of hydrogen and that hydrogen can be stored for later use in the fuel cell (Agbossou, 2004). The operation diagram of a hydrogen fuel cell energy storage system is presented in Figure 11. Hydrogen volatility and its atoms reduced dimension put the hydrogen storage reservoir as the critical element in this device. Last research place Metallic Hydrates as one of most efficient (Ogden, 1999). In the last years, hydrogen fuel cell systems become one of the most referred storage technologies to set up renewable energy integration issue. Price and charge/discharge efficiency about the 30% are its main constraints. Fig. 11. Hydrogen fuel cell energy storage system operation diagram 4. Energy Storage System Design The energy storage system design process consists off the determination of the storage power and energy capacity and the technologies that allow a better integration of the renewable sources of energy and the minimization of the thermal units fuel consumption and greenhouse effect gaseous emissions. The energy storage system design process is divided in two different phases. The first phase consists of the implementation of an unit commitment, including the energy storage system, in order to enable the technical evaluation of several power and energy storage capacity combinations and optimize their operation. Besides the operation optimization and the feasibility evaluation of each power and energy storage capacity combination, in the first phase of the design process are also determined the needs for renewable energy curtailment and the total thermal power units operation cost. The second phase of the energy storage system design process is based on an economical evaluation, where the costs and benefits associated to each technically feasible power and energy storage capacity combination are considered and the best techno-economical energy storage solution is determined. 4.1 Optimization of the energy storage system operation The energy storage system operation, along a time-series, is determined throughout an optimization process that manages the system in order to best integrate the renewable energy generation, allowing, consequently, a minimization of the thermal power plants costs. The thermal power plants costs are computed by adding up the fuel costs with the emission costs due to CO2. The optimization process is based on a power plant commitment problem, where the load, the renewable generation and the interconnections with other power grids are considered as input data, forecasted in a previous process. The energy storage system is integrated and operated as an additional power generation unit. The specificity of that unit is related to its ability to absorb power, especially in moments when the power system has no capacity to Embedded Energy Storage Systems in the Power Grid for Renewable Energy Sources Integration 73Renewable Energy72 Over the past few years three types of redox-flow batteries had been developed up to the stage of demonstration and commercialization. These types are vanadium redox batteries (VRB), the polysulphide bromide batteries (PSB) and the zinc bromine (ZnBr). (c) Compressed air Compressed air energy storage is a device based on as gas turbine where the compression and the combustion processes are divided. During charging, the compressor is coupled to the electrical machine, working as a motor, compressing the air. After the compression, the air is stored into a sealed underground cavern. Discharging the device consists in generating power through the coupling of the gas turbine with the electrical machine, working as generator, and supplying the stored compressed air to the combustion process (Lerch, 2007). A compressed air energy storage system operation diagram is presented in Figure 9. Fig. 9. Compressed air energy storage system operation diagram Three air reservoir types are generally considered: naturally occurring aquifers (such as those used for natural gas storage), solution-mined salt caverns, and mechanically formed reservoirs in rock formations. Main compressed air energy storage system implementation constraints are related with reservoirs achievement (Schoenung, 1996) (d) Hydrogen fuel cell A fuel cell is an energy conversion device that is closely related to a battery. Both are electrochemical devices for the conversion of chemical to electrical energy. In a battery the chemical energy is stored internally, whereas in a fuel cell the chemical energy (fuel and oxidant) is supplied externally and can be continuously replenished (Hoogers, 2003). The overall reaction in a fuel cell is the spontaneous reaction of hydrogen and oxygen to produce electricity and water. During the operation of a fuel cell, hydrogen is ionized into protons and electrons at the anode, the hydrogen ions are transported through the electrolyte to the cathode by an external circuit (load). At the cathode, oxygen combines with the hydrogen ions and electrons to produce water. The hydrogen fuel cell system can be reversible, allowing electric power consumption for the production of hydrogen and that hydrogen can be stored for later use in the fuel cell (Agbossou, 2004). The operation diagram of a hydrogen fuel cell energy storage system is presented in Figure 11. Hydrogen volatility and its atoms reduced dimension put the hydrogen storage reservoir as the critical element in this device. Last research place Metallic Hydrates as one of most efficient (Ogden, 1999). In the last years, hydrogen fuel cell systems become one of the most referred storage technologies to set up renewable energy integration issue. Price and charge/discharge efficiency about the 30% are its main constraints. Fig. 11. Hydrogen fuel cell energy storage system operation diagram 4. Energy Storage System Design The energy storage system design process consists off the determination of the storage power and energy capacity and the technologies that allow a better integration of the renewable sources of energy and the minimization of the thermal units fuel consumption and greenhouse effect gaseous emissions. The energy storage system design process is divided in two different phases. The first phase consists of the implementation of an unit commitment, including the energy storage system, in order to enable the technical evaluation of several power and energy storage capacity combinations and optimize their operation. Besides the operation optimization and the feasibility evaluation of each power and energy storage capacity combination, in the first phase of the design process are also determined the needs for renewable energy curtailment and the total thermal power units operation cost. The second phase of the energy storage system design process is based on an economical evaluation, where the costs and benefits associated to each technically feasible power and energy storage capacity combination are considered and the best techno-economical energy storage solution is determined. 4.1 Optimization of the energy storage system operation The energy storage system operation, along a time-series, is determined throughout an optimization process that manages the system in order to best integrate the renewable energy generation, allowing, consequently, a minimization of the thermal power plants costs. The thermal power plants costs are computed by adding up the fuel costs with the emission costs due to CO2. The optimization process is based on a power plant commitment problem, where the load, the renewable generation and the interconnections with other power grids are considered as input data, forecasted in a previous process. The energy storage system is integrated and operated as an additional power generation unit. The specificity of that unit is related to its ability to absorb power, especially in moments when the power system has no capacity to Embedded Energy Storage Systems in the Power Grid for Renewable Energy Sources Integration 73Renewable Energy74 accommodate all the renewable generation. A negative power of that power generation unit corresponds to charging the storage system and a positive power corresponds to its discharge. The power and energy limits of the energy storage system are imposed and considered as an input of the optimization problem. Therefore, it may be possible that, for some moments of the time series, the need for power storage overloads the storage capacity, and renewable power generation must be curtailed. In order to quantify the global renewable power and energy curtailed, a power unit that represents the excess of renewable generation is considered. The main objective of the energy storage system implementation is to support the integration of the renewable generation, being available to that integration as long as possible. For that reason, a penalty on the cost function has been introduced for the use of the storage system power charging, avoiding the temptation of the optimization process to exploit that system to attain the cheapest thermal power plants operation, when their minimal technical limits do not allow it. The penalty on the cost function does not affect the energy storage system negative power, enabling the system to be discharged anytime it is necessary or advantageous to avoid thermal power generation costs. The curtailment power has also a penalty in the cost function, greater than the energy storage system charging one, enabling its use only when it is strictly necessary. The objective function that is intended to be minimized considers the penalties presented above and the variable costs of the thermal power units, corresponding to their fuel costs and externalities resulting from the greenhouse effect gaseous emissions. Next, one presents the objective function:        TtJjjjVCtmCtmESSESS tPτCtPCtPCδz1 11:min (2) where T is the total time series period, τ is the time interval between t and t+1, J is the number of thermal power units,  tP j is the power generated by the thermal power unit j at moment t, jVC is the variable cost of generated energy of the thermal power unit j,  tPCtm is the renewable power curtailment at moment t, CtmC is the renewable power curtailment penalty  0 ,  tPESS is the power of the energy storage system at moment t, ESSC is the power charging penalty of the storage system  0 , finally 1δ is used to distinguish positive and negative energy storage system power:   0,10,01 tPtPδESSESS The balance between the power demand and the power generation is maintained by the following equation:            tQtLltPtPtP AVJjjCtmESS  11 (3) where  tL is the load power at moment t,  tQ is the sum of the renewable and the interconnections power and AVl is the average value of the losses in the power network. The commitment of the power units must consider the spinning reserve as presented in the next restriction:        tQtLltvPP AVJjjjMaxESSMax 11 (4) In (4) jMaxP is the maximum power generated by the thermal power unit j,  tv j is a binary variable that indicates, at moment t, if the thermal power unit j is running or if it is inactive. The term ESSMaxP corresponds to the maximum discharging power of the energy storage system. Considering the increasing integration of the wind power generation on the power systems and taking into account the volatility of that renewable energy source, there are some restrictions, introduced by the power systems operators, to the share of the wind power in the load power satisfaction. Therefore, in the moments when the wind power generated overcomes the level considered safe by the power system operator, the excess of wind power must be stored and the thermal power generation must be increased in order to fit the power demand. The following restriction reflects that concern:             tWtQtLlwδtP AVMaxJjj 1121 (5) where Maxw is the maximum share of wind power admitted by the power system operator,  tW is wind power generated at moment t, and 2δ distinguish the moments when the maximum share of wind power is overloaded:           tLlwtWtLlwtWδAVMaxAVMax1,11,02 In power system generation scheduling, the reflection of actual operating processes needs the use of ramp-rate constraints to simulate the thermal unit generation changes (Wang & Shahidepour, 1993). The thermal power units ramp-rates are considered and imposed by restrictions (6) and (7), corresponding to the ramp-up power rate and to the ramp-down power rate, respectively. No ramp-rates had been considered for the ESS.     jupjj PtPtP  1 (6)     jdownjj PtPtP 1 (7) where jupP and jdownP correspond to the ramp-up power rate and to the ramp-down power rate of the thermal power unit j. The limits to maximum and minimal power generation of each power unit are restricted by the conditions below:    tvPtP jjMaxj  (8) Embedded Energy Storage Systems in the Power Grid for Renewable Energy Sources Integration 75Renewable Energy74 accommodate all the renewable generation. A negative power of that power generation unit corresponds to charging the storage system and a positive power corresponds to its discharge. The power and energy limits of the energy storage system are imposed and considered as an input of the optimization problem. Therefore, it may be possible that, for some moments of the time series, the need for power storage overloads the storage capacity, and renewable power generation must be curtailed. In order to quantify the global renewable power and energy curtailed, a power unit that represents the excess of renewable generation is considered. The main objective of the energy storage system implementation is to support the integration of the renewable generation, being available to that integration as long as possible. For that reason, a penalty on the cost function has been introduced for the use of the storage system power charging, avoiding the temptation of the optimization process to exploit that system to attain the cheapest thermal power plants operation, when their minimal technical limits do not allow it. The penalty on the cost function does not affect the energy storage system negative power, enabling the system to be discharged anytime it is necessary or advantageous to avoid thermal power generation costs. The curtailment power has also a penalty in the cost function, greater than the energy storage system charging one, enabling its use only when it is strictly necessary. The objective function that is intended to be minimized considers the penalties presented above and the variable costs of the thermal power units, corresponding to their fuel costs and externalities resulting from the greenhouse effect gaseous emissions. Next, one presents the objective function:        TtJjjjVCtmCtmESSESS tPτCtPCtPCδz1 11:min (2) where T is the total time series period, τ is the time interval between t and t+1, J is the number of thermal power units,  tP j is the power generated by the thermal power unit j at moment t, jVC is the variable cost of generated energy of the thermal power unit j,  tPCtm is the renewable power curtailment at moment t, CtmC is the renewable power curtailment penalty  0 ,  tPESS is the power of the energy storage system at moment t, ESSC is the power charging penalty of the storage system  0 , finally 1δ is used to distinguish positive and negative energy storage system power:   0,10,01 tPtPδESSESS The balance between the power demand and the power generation is maintained by the following equation:            tQtLltPtPtP AVJjjCtmESS  11 (3) where  tL is the load power at moment t,  tQ is the sum of the renewable and the interconnections power and AVl is the average value of the losses in the power network. The commitment of the power units must consider the spinning reserve as presented in the next restriction:        tQtLltvPP AVJjjjMaxESSMax 11 (4) In (4) jMaxP is the maximum power generated by the thermal power unit j,  tv j is a binary variable that indicates, at moment t, if the thermal power unit j is running or if it is inactive. The term ESSMaxP corresponds to the maximum discharging power of the energy storage system. Considering the increasing integration of the wind power generation on the power systems and taking into account the volatility of that renewable energy source, there are some restrictions, introduced by the power systems operators, to the share of the wind power in the load power satisfaction. Therefore, in the moments when the wind power generated overcomes the level considered safe by the power system operator, the excess of wind power must be stored and the thermal power generation must be increased in order to fit the power demand. The following restriction reflects that concern:             tWtQtLlwδtP AVMaxJjj 1121 (5) where Maxw is the maximum share of wind power admitted by the power system operator,  tW is wind power generated at moment t, and 2δ distinguish the moments when the maximum share of wind power is overloaded:           tLlwtWtLlwtWδAVMaxAVMax1,11,02 In power system generation scheduling, the reflection of actual operating processes needs the use of ramp-rate constraints to simulate the thermal unit generation changes (Wang & Shahidepour, 1993). The thermal power units ramp-rates are considered and imposed by restrictions (6) and (7), corresponding to the ramp-up power rate and to the ramp-down power rate, respectively. No ramp-rates had been considered for the ESS.     jupjj PtPtP  1 (6)     jdownjj PtPtP 1 (7) where jupP and jdownP correspond to the ramp-up power rate and to the ramp-down power rate of the thermal power unit j. The limits to maximum and minimal power generation of each power unit are restricted by the conditions below:    tvPtP jjMaxj  (8) Embedded Energy Storage Systems in the Power Grid for Renewable Energy Sources Integration 75Renewable Energy76    tvPtP jjMinj  (9) In (9) jMinP corresponds to the technical minimum power that can be generated by the thermal power unit j. The energy storage system operation is limited by:   ESSMaxESS PtP  (10)   ESSMinESS PtP  (11) The ESSMaxP , as already referred, is the maximum discharging power of the storage system, and the ESSMinP , corresponds to the its maximum charging power (negative), or, alternatively, to the energy storage system minimum power. The renewable curtailment power intervention just makes sense when the energy storage system has no ability to absorb more renewable power generation. Therefore, as presented in (13), its value is never positive.   0tPCtm (13) As referred above, the storage system power and energy capacity are imposed and act like inputs of the problem. In the restrictions (10) and (11), the energy storage system power limits had been presented, and, next, one presents restrictions related to its stored energy limits. The stored energy at each moment of the time-series results from the integral of the energy storage system power in the preceding moments, and can be determined by the following equation:         ESSESSESSESSηδδtPτtEtE 111 11 (14) In (14)  tEESS is the stored energy in the storage system at moment t, and ESSη corresponds to the energy storage system charge/discharge power efficiency. The stored energy in the storage system is limited by restrictions (15) and (16).   ESSMaxESS EtE  (15)   ESSMinESS EtE  (16) where ESSMaxE and ESSMinE correspond to the maximum and minimum stored energy, respectively. Imposing that the initial state-of-charge of the energy storage system must be the same as at the end of period T, it comes:     011 111 ESSTtESSηδδtP (17) Once considered all the restrictions of the problem, the minimization of the objective function, because of the on-off solution for the thermal power units, can be provided by a mixed integer programming (Gollmer et al., 2000). 4.2 Cost of the energy storage system The costs associated with the acquisition of an energy storage system are much diversified and dependent on the technology adopted. The total energy storage system acquisition cost is composed by two different contributions; the contribution associated with peak power capacity of the storage device and the contribution associated to the amount of energy that can be stored. In Figure 12 a comparison of the acquisition costs for the commercialized energy storage technologies is presented. Fig. 12. Capital cost for the acquisition of different energy storage technologies (ESA, 2009) The long-term discharge technologies applicable at large storage systems, like pumping hydro, compressed air or flow batteries, present a low dependency between the energy storage capacity and the peak power. In those cases, the capital cost for the energy storage device acquisition can be expressed as the sum of the power capital cost with the energy capital cost (Barton & Infield, 2004 and Chacra et al., 2005): ESSdESSdESS EβPαC maxmax  (18) where dα is the capital cost per unit power of the technology d and dβ is the capital cost per unit energy of the technology d. Embedded Energy Storage Systems in the Power Grid for Renewable Energy Sources Integration 77Renewable Energy76    tvPtP jjMinj  (9) In (9) jMinP corresponds to the technical minimum power that can be generated by the thermal power unit j. The energy storage system operation is limited by:   ESSMaxESS PtP  (10)   ESSMinESS PtP  (11) The ESSMaxP , as already referred, is the maximum discharging power of the storage system, and the ESSMinP , corresponds to the its maximum charging power (negative), or, alternatively, to the energy storage system minimum power. The renewable curtailment power intervention just makes sense when the energy storage system has no ability to absorb more renewable power generation. Therefore, as presented in (13), its value is never positive.   0tPCtm (13) As referred above, the storage system power and energy capacity are imposed and act like inputs of the problem. In the restrictions (10) and (11), the energy storage system power limits had been presented, and, next, one presents restrictions related to its stored energy limits. The stored energy at each moment of the time-series results from the integral of the energy storage system power in the preceding moments, and can be determined by the following equation:         ESSESSESSESSηδδtPτtEtE 111 11 (14) In (14)  tEESS is the stored energy in the storage system at moment t, and ESSη corresponds to the energy storage system charge/discharge power efficiency. The stored energy in the storage system is limited by restrictions (15) and (16).   ESSMaxESS EtE  (15)   ESSMinESS EtE  (16) where ESSMaxE and ESSMinE correspond to the maximum and minimum stored energy, respectively. Imposing that the initial state-of-charge of the energy storage system must be the same as at the end of period T, it comes:     011 111 ESSTtESSηδδtP (17) Once considered all the restrictions of the problem, the minimization of the objective function, because of the on-off solution for the thermal power units, can be provided by a mixed integer programming (Gollmer et al., 2000). 4.2 Cost of the energy storage system The costs associated with the acquisition of an energy storage system are much diversified and dependent on the technology adopted. The total energy storage system acquisition cost is composed by two different contributions; the contribution associated with peak power capacity of the storage device and the contribution associated to the amount of energy that can be stored. In Figure 12 a comparison of the acquisition costs for the commercialized energy storage technologies is presented. Fig. 12. Capital cost for the acquisition of different energy storage technologies (ESA, 2009) The long-term discharge technologies applicable at large storage systems, like pumping hydro, compressed air or flow batteries, present a low dependency between the energy storage capacity and the peak power. In those cases, the capital cost for the energy storage device acquisition can be expressed as the sum of the power capital cost with the energy capital cost (Barton & Infield, 2004 and Chacra et al., 2005): ESSdESSdESS EβPαC maxmax  (18) where dα is the capital cost per unit power of the technology d and dβ is the capital cost per unit energy of the technology d. Embedded Energy Storage Systems in the Power Grid for Renewable Energy Sources Integration 77Renewable Energy78 Variable costs associated to the energy storage system operation and maintenance are not considered in the present work. Considering that the investment on the energy storage system acquisition can be amortized along its lifetime, it comes (Riggs et al., 1998):   NnnESSESSiAC1 1 (19) where ESSC is the initial investment on the energy storage system acquisition, n is one of the years from the total lifetime N, ESSA is the constant annual amortization of the initial investment and i is the interest rate. Developing the series (19), comes:         NNESSESS iiiiAC   1111 121  (20) Manipulating (20), the constant annuity ESSA can be obtained by:    111 NNESSESSiiiCA (21) Annuity ESSA should be computed for each energy storage solution achieved and used for its economical evaluation. 5. Location of the Energy Storage Systems in the Power Grid The implementation of the energy storage systems on the power grid allows a better management of the power flows due to the partial decoupling of the power demand and generation. Through the energy storage use it is possible to absorb the excess of renewable generation that occurs during the off-peak periods and inject that stored energy into the grid during peak hours, avoiding the operation of the most expensive and pollutant power thermal units. The carefully chosen location of the energy storage systems in the power grid can avoid non-desired power flows or congestions, improving the flexibility and the efficiency of the power grid. Following, three different perspectives about the power grid energy storage systems location are presented and discussed. 5.1 Distributed energy storage at generation The concept of distributed energy storage at generation considers that the energy storage systems are located near the most important non-dispatchable renewable power producers, such as the wind farms. Distributed energy storage at generation is the most intuitive perspective, because it mitigates the power unbalance issue at the original source. In fact, as far as energy management is concerned, the distributed energy storage at generation perspective is as suitable as any other approach. However, from this energy storage location, a considerable contribution in terms of global power grid congestions reduction is not expected. In this case, the energy storage systems are located near the renewable power generation and considering that the positive unbalance is prominent during off-peak hours, the stored energy is discharged into the system, leading to an increase of the power flows when the grid is eventually already congested. One of the advantages of the distributed energy storage at generation is related to the possibility of downsizing the equipment used for the connection of the renewable producers to the power grid. This is due to the contribution of the energy storage system for the modulation of the renewable generation shape, enabling a less variable power output and avoiding the power peaks. The implementation of distributed energy storage at generation depends on the availability of modular long-term discharge energy storage devices adapted to the specific characteristics of each renewable power plant (Faias et al., 2008). 5.2 Distributed energy storage at demand Distributed energy storage at demand is a concept that considers the energy storage devices located near the demand, eventually in the interface between the transmission and distribution networks. Distributed energy storage at demand promotes power flows at off-peak periods and relieves the system from power flows at peak periods. As the energy storage devices are disposed close to the regions where the energy is going to be consumed, the power grid congestions can be avoided, because the power flows between the renewable energy producers and the energy storage system is going to occur mainly during the off-peak periods. During peak hours, when the energy storage system is discharged, the power loads are partially supplied by that power and the needs for high power flows on the transmission network will be reduced. Like the distributed energy storage at generation concept, the distributed energy storage at demand concept implies the use of modular medium-size long-term discharge energy storage devices. 5.3 Centralized energy storage The concept of centralized energy storage is based on the idea of installing some (few) large energy storage devices on the power system in order to better manage the power unbalances induced by the increasing integration of the renewable power sources. It is expected that the energy storage, being centralized in a small number of devices, does not contribute for an improvement on the power flows and congestions for all power grid topologies, because the power flows resulting from the excess of renewable generation will converge to the places where the centralized energy storage devices are located. In addition, during the discharge periods, the power flows on the energy storage devices neighbourhood are also going to rise, increasing the congestions and resulting in a potential negative contribution for the overall power system efficiency. The main centralized energy storage advantages come from the energy storage technologies that can be applied as far as this concept is concerned. Large-size long-term discharge energy storage devices like pumping-hydro and compressed air are best suited for this kind of centralized application. These technologies present the lower capital costs per power and Embedded Energy Storage Systems in the Power Grid for Renewable Energy Sources Integration 79Renewable Energy78 Variable costs associated to the energy storage system operation and maintenance are not considered in the present work. Considering that the investment on the energy storage system acquisition can be amortized along its lifetime, it comes (Riggs et al., 1998):   NnnESSESSiAC1 1 (19) where ESSC is the initial investment on the energy storage system acquisition, n is one of the years from the total lifetime N, ESSA is the constant annual amortization of the initial investment and i is the interest rate. Developing the series (19), comes:         NNESSESS iiiiAC   1111 121  (20) Manipulating (20), the constant annuity ESSA can be obtained by:    111 NNESSESSiiiCA (21) Annuity ESSA should be computed for each energy storage solution achieved and used for its economical evaluation. 5. Location of the Energy Storage Systems in the Power Grid The implementation of the energy storage systems on the power grid allows a better management of the power flows due to the partial decoupling of the power demand and generation. Through the energy storage use it is possible to absorb the excess of renewable generation that occurs during the off-peak periods and inject that stored energy into the grid during peak hours, avoiding the operation of the most expensive and pollutant power thermal units. The carefully chosen location of the energy storage systems in the power grid can avoid non-desired power flows or congestions, improving the flexibility and the efficiency of the power grid. Following, three different perspectives about the power grid energy storage systems location are presented and discussed. 5.1 Distributed energy storage at generation The concept of distributed energy storage at generation considers that the energy storage systems are located near the most important non-dispatchable renewable power producers, such as the wind farms. Distributed energy storage at generation is the most intuitive perspective, because it mitigates the power unbalance issue at the original source. In fact, as far as energy management is concerned, the distributed energy storage at generation perspective is as suitable as any other approach. However, from this energy storage location, a considerable contribution in terms of global power grid congestions reduction is not expected. In this case, the energy storage systems are located near the renewable power generation and considering that the positive unbalance is prominent during off-peak hours, the stored energy is discharged into the system, leading to an increase of the power flows when the grid is eventually already congested. One of the advantages of the distributed energy storage at generation is related to the possibility of downsizing the equipment used for the connection of the renewable producers to the power grid. This is due to the contribution of the energy storage system for the modulation of the renewable generation shape, enabling a less variable power output and avoiding the power peaks. The implementation of distributed energy storage at generation depends on the availability of modular long-term discharge energy storage devices adapted to the specific characteristics of each renewable power plant (Faias et al., 2008). 5.2 Distributed energy storage at demand Distributed energy storage at demand is a concept that considers the energy storage devices located near the demand, eventually in the interface between the transmission and distribution networks. Distributed energy storage at demand promotes power flows at off-peak periods and relieves the system from power flows at peak periods. As the energy storage devices are disposed close to the regions where the energy is going to be consumed, the power grid congestions can be avoided, because the power flows between the renewable energy producers and the energy storage system is going to occur mainly during the off-peak periods. During peak hours, when the energy storage system is discharged, the power loads are partially supplied by that power and the needs for high power flows on the transmission network will be reduced. Like the distributed energy storage at generation concept, the distributed energy storage at demand concept implies the use of modular medium-size long-term discharge energy storage devices. 5.3 Centralized energy storage The concept of centralized energy storage is based on the idea of installing some (few) large energy storage devices on the power system in order to better manage the power unbalances induced by the increasing integration of the renewable power sources. It is expected that the energy storage, being centralized in a small number of devices, does not contribute for an improvement on the power flows and congestions for all power grid topologies, because the power flows resulting from the excess of renewable generation will converge to the places where the centralized energy storage devices are located. In addition, during the discharge periods, the power flows on the energy storage devices neighbourhood are also going to rise, increasing the congestions and resulting in a potential negative contribution for the overall power system efficiency. The main centralized energy storage advantages come from the energy storage technologies that can be applied as far as this concept is concerned. Large-size long-term discharge energy storage devices like pumping-hydro and compressed air are best suited for this kind of centralized application. These technologies present the lower capital costs per power and Embedded Energy Storage Systems in the Power Grid for Renewable Energy Sources Integration 79Renewable Energy80 energy units. Despite the geographical constraints associated to each one of these storage devices, as the number of units to be implemented is small, it is not expected that it would be much difficult to find some location presenting adequate physical conditions for their installation. 6. Case Study: Portuguese Power System at 2011 The case study hereby presented corresponds to the Portuguese power system, which has experienced an increasing integration of renewable power sources during the last years. The currently installed capacity in Portugal is about 15 GW and comprises coal (12%), natural gas (15%), fuel oil (11%) and diesel fuel utilities (1%), referred as thermal technologies; reservoir (16%) and run-of-river (14%), referred as hydro with more than 10 MW; cogeneration (6%), wind (18%), small hydro that account for less than 10 MW (2%), biomass (2%), municipal solid waste (0.6%) and photovoltaic (0.4%), referred as renewable (REN, 2008). 2008 was a year in which hydro generation presented a value 44% lower than the usual. That is why the renewable generation was only about 27% of the consumed energy. The main contributors for the renewable generation were the hydro power (12%), the wind power (11%) and the biomass (4%). For the year 2010, Portuguese authorities estimate that the renewable generation will correspond to 45% of the consumed energy. Wind power is the renewable source of energy that has experienced the highest increase in the Portuguese power system. At the end of the year 2008, the wind power installed capacity was about 2800 MW, 11 times higher than at the 2003. According to the national power system operator predictions, by year 2010, wind power will correspond to 25% of the installed capacity. As already referred in this chapter, such significant share of a variable and non-dispatchable renewable power source could induce some problems related to the power system unbalance. In the sequence, one presents an evaluation of the power unbalance issue in the Portuguese power system, by year 2011, and discusses possible solutions for that problem. 6.1 Power unbalance forecast The power unbalance forecast intends to evaluate the impact of the increasing integration of renewable power sources, namely wind power, in the power unbalance between generation and demand on the Portuguese power system by year 2011. The forecast procedure is based on the Portuguese power system historic data of 2007 by power plant and by aggregated demand buses at transmission level. Forecasts for the demand, renewable generation and interconnections with other power grids use the historic data adjusted to the predicted installed capacity and energy consumption by the year 2011. Thermal power generation predictions have been computed using the above mentioned (section 4.2) thermal unit commitment methodology, based on the power demand needs and considering the fuel and pollutant emissions costs. Thermal power spinning reserve was determined also through the historic data analysis. An average power grid losses value have been also considered and added to the power demand. Therefore, considering all conditions referred above, the power unbalance had been forecasted and the obtained results are presented in Figure 13. Forecast results show that the power generation will overcome the power demand during 25% of the year, corresponding to an energy equivalent to 9.5% of the annual wind power generation. It is also possible to verify that, in selected moments, the power unbalance reaches almost 50% of the annual average load. Fig. 13. Time-series and duration curve of the power unbalance forecast (year 2011) In order to better understand the hourly distribution of the power unbalance along the day, the equivalent average day of the forecasted annual power unbalance time-series was determined. The result is presented in Figure 14 where, as expected, the emphasis goes to the power unbalance prominence during the off-peak hours. This fact results from the time coincidence of the lowest power demand with the highest wind power availability. Fig. 14. Equivalent average day of the forecasted annual power unbalance Regarding the excess of power generation estimated by the forecast and the need to mitigate that power unbalance, two different, but complementary, solutions can be adopted. The first one corresponds to the implementation of an energy storage system able to absorb the excess of generation, and the second one, the curtailment of the wind power generation. Taking into account that the renewable power generation should not be wasted because of its several benefits, the wind power curtailment must be assumed as the last option. Embedded Energy Storage Systems in the Power Grid for Renewable Energy Sources Integration 81Renewable Energy80 energy units. Despite the geographical constraints associated to each one of these storage devices, as the number of units to be implemented is small, it is not expected that it would be much difficult to find some location presenting adequate physical conditions for their installation. 6. Case Study: Portuguese Power System at 2011 The case study hereby presented corresponds to the Portuguese power system, which has experienced an increasing integration of renewable power sources during the last years. The currently installed capacity in Portugal is about 15 GW and comprises coal (12%), natural gas (15%), fuel oil (11%) and diesel fuel utilities (1%), referred as thermal technologies; reservoir (16%) and run-of-river (14%), referred as hydro with more than 10 MW; cogeneration (6%), wind (18%), small hydro that account for less than 10 MW (2%), biomass (2%), municipal solid waste (0.6%) and photovoltaic (0.4%), referred as renewable (REN, 2008). 2008 was a year in which hydro generation presented a value 44% lower than the usual. That is why the renewable generation was only about 27% of the consumed energy. The main contributors for the renewable generation were the hydro power (12%), the wind power (11%) and the biomass (4%). For the year 2010, Portuguese authorities estimate that the renewable generation will correspond to 45% of the consumed energy. Wind power is the renewable source of energy that has experienced the highest increase in the Portuguese power system. At the end of the year 2008, the wind power installed capacity was about 2800 MW, 11 times higher than at the 2003. According to the national power system operator predictions, by year 2010, wind power will correspond to 25% of the installed capacity. As already referred in this chapter, such significant share of a variable and non-dispatchable renewable power source could induce some problems related to the power system unbalance. In the sequence, one presents an evaluation of the power unbalance issue in the Portuguese power system, by year 2011, and discusses possible solutions for that problem. 6.1 Power unbalance forecast The power unbalance forecast intends to evaluate the impact of the increasing integration of renewable power sources, namely wind power, in the power unbalance between generation and demand on the Portuguese power system by year 2011. The forecast procedure is based on the Portuguese power system historic data of 2007 by power plant and by aggregated demand buses at transmission level. Forecasts for the demand, renewable generation and interconnections with other power grids use the historic data adjusted to the predicted installed capacity and energy consumption by the year 2011. Thermal power generation predictions have been computed using the above mentioned (section 4.2) thermal unit commitment methodology, based on the power demand needs and considering the fuel and pollutant emissions costs. Thermal power spinning reserve was determined also through the historic data analysis. An average power grid losses value have been also considered and added to the power demand. Therefore, considering all conditions referred above, the power unbalance had been forecasted and the obtained results are presented in Figure 13. Forecast results show that the power generation will overcome the power demand during 25% of the year, corresponding to an energy equivalent to 9.5% of the annual wind power generation. It is also possible to verify that, in selected moments, the power unbalance reaches almost 50% of the annual average load. Fig. 13. Time-series and duration curve of the power unbalance forecast (year 2011) In order to better understand the hourly distribution of the power unbalance along the day, the equivalent average day of the forecasted annual power unbalance time-series was determined. The result is presented in Figure 14 where, as expected, the emphasis goes to the power unbalance prominence during the off-peak hours. This fact results from the time coincidence of the lowest power demand with the highest wind power availability. Fig. 14. Equivalent average day of the forecasted annual power unbalance Regarding the excess of power generation estimated by the forecast and the need to mitigate that power unbalance, two different, but complementary, solutions can be adopted. The first one corresponds to the implementation of an energy storage system able to absorb the excess of generation, and the second one, the curtailment of the wind power generation. Taking into account that the renewable power generation should not be wasted because of its several benefits, the wind power curtailment must be assumed as the last option. Embedded Energy Storage Systems in the Power Grid for Renewable Energy Sources Integration 81Renewable Energy82 The use of the interconnections with the Spanish power system is not considered as a solution to mitigate the power unbalance. Spanish power system also presents an important share of renewable energy, namely wind power, and probably the power unbalance issue affects similarly both power systems. Moreover, historic interconnection data shows that during the off-peak periods, when the forecasted power unbalance tends to be more intense, the fluxes of energy are typically from Spain to Portugal. 6.2 Energy storage system design The energy storage system design follows the procedure previously described in section 4. Different power and energy combinations for the energy storage system are introduced in the unit commitment problem and their feasibility is evaluated. The economical benefit that result from the cost reduction of the fuel consumption and greenhouse gases emissions due to the lower renewable power curtailment are also computed during that procedure. Results are presented in Figure 15. a) Costs reduction b) Renewable curtailment Fig. 15. Impact of different power and energy storage combinations on the reduction of annual thermal power plants costs and on the curtailment of the renewable generation The economical benefit resulting from the reduction of fuel consumption and greenhouse gases emissions due to the implementation of the energy storage system is directly proportional to the increment on the storage power and energy capacity, while it is inversely proportional to the need of renewable generation curtailment. The energy storage system able to completely match the power unbalance would need a global capacity of 3270 MW and 395 GWh. Together with the thermal power plants economical benefits induced by the energy storage system, also the capital costs associated with the energy storage system acquisition must be considered. Therefore, the capital costs of the different technologies are deduced to the benefits that result from their implementation. The net economic benefit was computed for the long-term discharge energy storage technologies like compressed air, pumping hydro and redox-flow batteries, assuming the interval between the maximum and minimum capital cost of the energy storage system acquisition (Figure 12) and an interest rate of 5%. The lifetime assumed for each technology was respectively 30 years, 40 years and 15 years (Schoenung et al., 1996) (EPRI, 2007). Hydrogen fuel cell energy storage technology was not considered in this procedure due to its low charge/discharge efficiency. From the three technologies considered, it was concluded that only compressed air energy storage presents some positive net economic benefits, when assuming the minimal value of the acquisition cost interval. The best economic result corresponds to an annual benefit of 1.2 million euros for a 100 MW and 1500 MWh storage system. Despite the contribution of this storage system to the power unbalance mitigation, its small capacity forces a renewable generation curtailment equivalent to 8.6% of the annual wind power generation. The solution that presents the second best economic result corresponds to a 50 MW and 500 MWh energy storage system and allows an annual benefit of 690000 euros. The renewable generation curtailment corresponding to this system is equivalent to 9% of the annual wind power generation. An energy storage system of 200 MW and 5000 MWh capacity presents the third best economic result, allowing a benefit of 570000 euros and a renewable generation curtailment equivalent to 7.7% of the annual wind power generation. This system, from all the energy storage solutions with positive economic results, is the one that better absorb the power unbalance. The 200 MW and 5000 MWh energy storage system operation along the period of one typical week, and its contribution for balancing the power generation and demand is presented in Figure 16. Fig. 16. Energy storage system (200 MW and 5000 MWh) operation and its contribution for balancing the power generation and demand along the period of one week Through the analysis of the figure, it is possible to identify, during the off-peak hours, the moments when the energy storage is charging, in order to absorb the excess of power generation. Also the energy storage system discharging moments are easily identified, during the peak hours, when the energy storage system is contributing to avoid additional thermal power generation. Embedded Energy Storage Systems in the Power Grid for Renewable Energy Sources Integration 83Renewable Energy82 The use of the interconnections with the Spanish power system is not considered as a solution to mitigate the power unbalance. Spanish power system also presents an important share of renewable energy, namely wind power, and probably the power unbalance issue affects similarly both power systems. Moreover, historic interconnection data shows that during the off-peak periods, when the forecasted power unbalance tends to be more intense, the fluxes of energy are typically from Spain to Portugal. 6.2 Energy storage system design The energy storage system design follows the procedure previously described in section 4. Different power and energy combinations for the energy storage system are introduced in the unit commitment problem and their feasibility is evaluated. The economical benefit that result from the cost reduction of the fuel consumption and greenhouse gases emissions due to the lower renewable power curtailment are also computed during that procedure. Results are presented in Figure 15. a) Costs reduction b) Renewable curtailment Fig. 15. Impact of different power and energy storage combinations on the reduction of annual thermal power plants costs and on the curtailment of the renewable generation The economical benefit resulting from the reduction of fuel consumption and greenhouse gases emissions due to the implementation of the energy storage system is directly proportional to the increment on the storage power and energy capacity, while it is inversely proportional to the need of renewable generation curtailment. The energy storage system able to completely match the power unbalance would need a global capacity of 3270 MW and 395 GWh. Together with the thermal power plants economical benefits induced by the energy storage system, also the capital costs associated with the energy storage system acquisition must be considered. Therefore, the capital costs of the different technologies are deduced to the benefits that result from their implementation. The net economic benefit was computed for the long-term discharge energy storage technologies like compressed air, pumping hydro and redox-flow batteries, assuming the interval between the maximum and minimum capital cost of the energy storage system acquisition (Figure 12) and an interest rate of 5%. The lifetime assumed for each technology was respectively 30 years, 40 years and 15 years (Schoenung et al., 1996) (EPRI, 2007). Hydrogen fuel cell energy storage technology was not considered in this procedure due to its low charge/discharge efficiency. From the three technologies considered, it was concluded that only compressed air energy storage presents some positive net economic benefits, when assuming the minimal value of the acquisition cost interval. The best economic result corresponds to an annual benefit of 1.2 million euros for a 100 MW and 1500 MWh storage system. Despite the contribution of this storage system to the power unbalance mitigation, its small capacity forces a renewable generation curtailment equivalent to 8.6% of the annual wind power generation. The solution that presents the second best economic result corresponds to a 50 MW and 500 MWh energy storage system and allows an annual benefit of 690000 euros. The renewable generation curtailment corresponding to this system is equivalent to 9% of the annual wind power generation. An energy storage system of 200 MW and 5000 MWh capacity presents the third best economic result, allowing a benefit of 570000 euros and a renewable generation curtailment equivalent to 7.7% of the annual wind power generation. This system, from all the energy storage solutions with positive economic results, is the one that better absorb the power unbalance. The 200 MW and 5000 MWh energy storage system operation along the period of one typical week, and its contribution for balancing the power generation and demand is presented in Figure 16. Fig. 16. Energy storage system (200 MW and 5000 MWh) operation and its contribution for balancing the power generation and demand along the period of one week Through the analysis of the figure, it is possible to identify, during the off-peak hours, the moments when the energy storage is charging, in order to absorb the excess of power generation. Also the energy storage system discharging moments are easily identified, during the peak hours, when the energy storage system is contributing to avoid additional thermal power generation. Embedded Energy Storage Systems in the Power Grid for Renewable Energy Sources Integration 83Renewable Energy84 The results obtained from the energy storage system design procedure, applied to the Portuguese case study, lead to the conclusion that the economically viable solutions only enable the integration of a small percentage of the power unbalance induced by the renewable generation. However, it should be emphasized that assuming higher fuel and greenhouse gases emissions costs, as well as lower storage systems capital costs, the economic viability of larger integration of the power unbalance would be improved. 6.3 Power system with integrated pumping hydro Some power systems with large amount of hydro installed capacity, such as the Portuguese system, comprise reversible hydro power plants with energy storage capacity that are usually used for the energy management improvement. In the present scenario it is assumed that all of the pumping hydro installed capacity is available for helping the integration of the renewable power generation (in opposition to the previous scenario where the pumping hydro system was operated according to the historic data). The estimated Portuguese pumping hydro capacity corresponds to 1051 MW and 25.2 GWh, and as far as the power unbalance forecast evaluation is concerned, it had been introduced in unit commitment problem as an additional energy storage system. As the Portuguese pumping hydro power plants are integrated in river flows, besides the usual constraints considered for all energy storage systems, an additional constraint was introduced to the pumping hydro operation, in order to consider the forced power discharges that result from the predicted up-side reservoir water inflows. Therefore, the power unbalance forecast procedure had been applied assuming that the pumping hydro system is completely available to cooperate in the integration of the renewable generation. As expected, that availability results in a lower power unbalance, reducing the renewable curtailment need for a value equivalent to 4% of the annual wind power generation. In order to evaluate the viability of an additional energy storage capacity for reducing the need for renewable power curtailment, the energy storage design process was again applied. Results show that, in this scenario, there is only one economically viable power and energy storage capacity combination. The energy storage system corresponds to a single compressed air energy storage unit with 10 MW, and 100 MWh. The annual net economic benefit of the energy storage system implementation is 780000 euros, and it is only viable assuming the minimum capital costs for the compressed air energy storage technology. The benefit introduced by this storage system, in terms of the integration of renewable power generation, is marginal. The renewable power generation curtailment, before the implementation of the energy storage system, was equivalent to 4% of the annual wind power generation, and after that it was 3.9%. One of the objectives of the design process economical evaluation is to identify a particular energy storage system capacity and technology, however, the specificities of that technology should be carefully considered in order to validate its technical feasibility. In Figure 17, where contribution of the two different energy storage systems for the power demand and generation balancing is presented, it is possible to follow the desired response for the compressed air energy storage system, referred as the other storage. Remembering that the compressed energy storage system operation is based on a gas turbine, difficultly it will respond to such demanding operation conditions. Fig. 17. Pumping hydro and the designed energy storage system operation and their contribution for balancing the power generation and demand along the period of one week 7. Conclusion In the last years the world has assisted to an impressing development of the renewable energies, driven by the need for reducing negative environmental impacts that resulted from an intense exploitation of fossil fuels. Besides the environmental concern, as the renewable energy sources are endogenous to many countries, also the security of supply has been one of the key factors responsible for that development of the renewable energy. The electricity generation sector, in particular, is the one that more intensely has experienced the replacement of fossil fuel based technologies by technologies that make use of renewable energy sources. As a result of that change, new challenges and issues have emerged on the operation of the power systems. As most of renewable technologies ultimately derive energy from unpredictable natural sources or sources that vary in their availability over different timescales, balancing power generation and demand is one of the issues that is raised with an increasing integration of renewable energy in the power system. Balancing power generation and demand is an issue that occurs especially during the off-peak periods, when typically the power demand is lower and the non-dispatchable renewable power can become higher than the demand. There are three different approaches to mitigate the power unbalance resulting from an excess of renewable power generation; curtailment of the renewable generation, interconnections with other grids and the energy storage. However, in order to avoid the waste of renewable generation and considering that different interconnected power systems Embedded Energy Storage Systems in the Power Grid for Renewable Energy Sources Integration 85Renewable Energy84 The results obtained from the energy storage system design procedure, applied to the Portuguese case study, lead to the conclusion that the economically viable solutions only enable the integration of a small percentage of the power unbalance induced by the renewable generation. However, it should be emphasized that assuming higher fuel and greenhouse gases emissions costs, as well as lower storage systems capital costs, the economic viability of larger integration of the power unbalance would be improved. 6.3 Power system with integrated pumping hydro Some power systems with large amount of hydro installed capacity, such as the Portuguese system, comprise reversible hydro power plants with energy storage capacity that are usually used for the energy management improvement. In the present scenario it is assumed that all of the pumping hydro installed capacity is available for helping the integration of the renewable power generation (in opposition to the previous scenario where the pumping hydro system was operated according to the historic data). The estimated Portuguese pumping hydro capacity corresponds to 1051 MW and 25.2 GWh, and as far as the power unbalance forecast evaluation is concerned, it had been introduced in unit commitment problem as an additional energy storage system. As the Portuguese pumping hydro power plants are integrated in river flows, besides the usual constraints considered for all energy storage systems, an additional constraint was introduced to the pumping hydro operation, in order to consider the forced power discharges that result from the predicted up-side reservoir water inflows. Therefore, the power unbalance forecast procedure had been applied assuming that the pumping hydro system is completely available to cooperate in the integration of the renewable generation. As expected, that availability results in a lower power unbalance, reducing the renewable curtailment need for a value equivalent to 4% of the annual wind power generation. In order to evaluate the viability of an additional energy storage capacity for reducing the need for renewable power curtailment, the energy storage design process was again applied. Results show that, in this scenario, there is only one economically viable power and energy storage capacity combination. The energy storage system corresponds to a single compressed air energy storage unit with 10 MW, and 100 MWh. The annual net economic benefit of the energy storage system implementation is 780000 euros, and it is only viable assuming the minimum capital costs for the compressed air energy storage technology. The benefit introduced by this storage system, in terms of the integration of renewable power generation, is marginal. The renewable power generation curtailment, before the implementation of the energy storage system, was equivalent to 4% of the annual wind power generation, and after that it was 3.9%. One of the objectives of the design process economical evaluation is to identify a particular energy storage system capacity and technology, however, the specificities of that technology should be carefully considered in order to validate its technical feasibility. In Figure 17, where contribution of the two different energy storage systems for the power demand and generation balancing is presented, it is possible to follow the desired response for the compressed air energy storage system, referred as the other storage. Remembering that the compressed energy storage system operation is based on a gas turbine, difficultly it will respond to such demanding operation conditions. Fig. 17. Pumping hydro and the designed energy storage system operation and their contribution for balancing the power generation and demand along the period of one week 7. Conclusion In the last years the world has assisted to an impressing development of the renewable energies, driven by the need for reducing negative environmental impacts that resulted from an intense exploitation of fossil fuels. Besides the environmental concern, as the renewable energy sources are endogenous to many countries, also the security of supply has been one of the key factors responsible for that development of the renewable energy. The electricity generation sector, in particular, is the one that more intensely has experienced the replacement of fossil fuel based technologies by technologies that make use of renewable energy sources. As a result of that change, new challenges and issues have emerged on the operation of the power systems. As most of renewable technologies ultimately derive energy from unpredictable natural sources or sources that vary in their availability over different timescales, balancing power generation and demand is one of the issues that is raised with an increasing integration of renewable energy in the power system. Balancing power generation and demand is an issue that occurs especially during the off-peak periods, when typically the power demand is lower and the non-dispatchable renewable power can become higher than the demand. There are three different approaches to mitigate the power unbalance resulting from an excess of renewable power generation; curtailment of the renewable generation, interconnections with other grids and the energy storage. However, in order to avoid the waste of renewable generation and considering that different interconnected power systems Embedded Energy Storage Systems in the Power Grid for Renewable Energy Sources Integration 85Renewable Energy86 can be simultaneously with similar difficulties for balancing their systems, the use of an energy storage system seems to be the best suited solution. At the moment, there are several energy storage technologies available for helping the integration of renewable energy sources in the power system. From those technologies one distinguishes among the pumping hydro energy storage, the compressed air energy storage and redox-flow batteries. Different technologies correspond to different power and energy capacity ranges and to different capital costs. Pumping hydro and compressed air energy storage systems present the larger power and energy capacity ranges and the lower investment costs. However, there are some restrictions to their implementation due to the need for specific geographical conditions. The modularity of the redox-flow batteries put great expectative in its use, nevertheless their investment costs make them few competitive at present. An energy storage system design methodology had been developed along the present chapter. The main goal of this design methodology was to choose an energy storage system able to integrate as much renewable generation as possible in the power system, taking into account the trade off between economic benefits and its capital costs. The first phase of the energy storage system design procedure consists of the evaluation of the economic benefit of the lower thermal power plants generation that results from the operation of several power and energy capacity storage combinations. During the second phase, the capital costs of the several power and energy capacity of the storage systems are computed and the best solutions are identified. In the case study presented for the Portuguese system it is showed that the implementation of the energy storage equipment in power systems with high penetration of renewable power sources, improves the power demand and generation balance and reduces the need for renewable generation curtailment. Some power systems with large amount of hydro installed capacity, such as the Portuguese system, comprise reversible hydro power plants with energy storage capacity, usually used for the energy management improvement. Therefore, two opposite scenarios were considered: the first scenario is related to the pumping hydro system operation accordingly to the historic data; in the second scenario it was assumed that all of the pumping hydro installed capacity was available for helping the integration of the renewable power generation. As far the energy storage systems implementation is concerned, those scenarios allowed the identification of their maximum and minimum capacity requirements. The availability of the installed pumping hydro systems, to cooperate in the integration of the renewable generation appears as a very positive solution for the power unbalance mitigation, especially when considering the costs of completely new energy storage systems implementation. The energy storage system design process needs then to consider the physical constraints of power grid. Therefore, in order to validate the results from that process, namely the power flows imposed by the unit commitment problem outputs, a power system simulator was used (PSS/E - Siemens) for the simulation of the Portuguese transmission network model. All the solutions were validated using the centralized energy storage concept. The solution that does not consider the pumping hydro availability had been validated also with the distributed energy storage at demand and the distributed energy storage at generation concepts. Regarding to the power grid losses, in this specific case study, the three different energy storage locations presented very similar results. 8. References Agbossou, K.; Kolhe, M.; Hamelin, J. & Bose, T. K. (2004). Performance of a Stand-Alone Renewable Energy System Based on Energy Storage as Hydrogen. IEEE Transactions on Energy Conversion, Vol. 19, No. 3, (Sept. 2004), page numbers (633-640), ISSN 0885-8969 Barton, J. P. & Infield, D. G. (2004). Energy Storage and Its Use with Intermittent Renewable Energy. IEEE Transactions on Energy Conversion, Vol. 19, No. 2, (June 2004), page numbers (441-448), ISSN 0885-8969 Chacra, F.; Bastard, P.; Fleury, G. & Clavreul, R. (2005). Impact of Energy Storage Costs on Economical Performance in a Distribution Substation. IEEE Transactions on Power Systems, Vol. 20, No. 2, (May 2005), page numbers (684-691), ISSN 0885-8950 EPRI (2007). Vanadium Redox Flow Batteries: An In-Depth Analysis. EPRI, Palo Alto, CA: 2007. 1014836 ESA (2009). Technology Comparisons, Electricity Storage Association (Online), Available: http://www.electricitystorage.org Faias, S.; Sousa, J.; Castro, R. (2007). Contribution of energy storage systems for power generation and demand balancing with increasing integration of renewable sources: application to the Portuguese power system, Proceedings of the 12th European Conference on Power Electronics and Applications, EPE 2007, pp. 1-10, Vol. 1, ISBN 978-92-75815-10-8, Aalborg, September 2007 Faias, S.; Santos, P.; Matos, F.; Sousa, J.; Castro, R. (2008). Evaluation of energy storage devices for renewable energies integration: Application to a Portuguese wind farm, Proceedings of the 5th International Conference on the European Electricity Market, EEM08, pp. 1-7, Vol. 1, ISBN 978-1-4244-1743-8, Lisbon, May 2008 Gollmer, R.; Nowak, M.; Römisch, W. & Schultz, R. (2000). Unit commitment in power generation – a basic model and some extensions. Annals of Operations Research, Vol. 96, No. 1-4, (November 2000) page numbers (167-189), ISSN 0254-5330 Hammons, T. J. (2006). Integrating Renewable Energy Sources into European Grids, Proceedings of the 41st International Universities Power Engineering Conference, pp. 142-151, Vol. 1, ISBN 978-186135-342-9, Newcastle upon Tyne, September 2006 Hamsic, N.; Schmelter, A.; Mohd, A.; Ortjohann, E.; Schultze, E.; Tuckey, A. & Zimmermann, J. (2007). Increasing Renewable Energy Penetration in Isolated Grids Using a Flywheel Energy Storage System, Proceedings of Powereng 2007 - International Conference on Power Engineering, Energy and Electrical Drives, pp. 195-200, ISBN 978-1-4244-0895-5, Setúbal – Portugal, April 2007 Hoogers, G. (2003). Fuel Cell Technology Handbook, CRC Press, ISBN 0-84-930877-1 Hsu, C. & Lee, W. (1992). Superconducting Magnetic Energy Storage for Power System Applications. IEEE Transactions on Industrial Applications, Vol. 29, No. 5, (Sept.-Oct. 1993), page numbers (990-996), ISSN 0093-9994 Hunt, G. L. (1998). The Great Battery Search. Spectrum IEEE, Vol. 35, No. 11, (Nov. 1998), page numbers (21-28), ISSN 0018-9235 IEA (2005). Variability of Wind Power and other Renewables: Management Options and Strategies. International Energy Agency Publications, Paris Lazarewicz, M. & Rojas, A. (2004). Grid Frequency Regulation by Recycling Electrical Energy in Flywheels, Proceedings of IEEE-Power Engineering Society General Meeting, Vol.2, pp. 2038- 2042, ISBN 0-7803-8465-2, Denver, CO, June 2004 Embedded Energy Storage Systems in the Power Grid for Renewable Energy Sources Integration 87Renewable Energy86 can be simultaneously with similar difficulties for balancing their systems, the use of an energy storage system seems to be the best suited solution. At the moment, there are several energy storage technologies available for helping the integration of renewable energy sources in the power system. From those technologies one distinguishes among the pumping hydro energy storage, the compressed air energy storage and redox-flow batteries. Different technologies correspond to different power and energy capacity ranges and to different capital costs. Pumping hydro and compressed air energy storage systems present the larger power and energy capacity ranges and the lower investment costs. However, there are some restrictions to their implementation due to the need for specific geographical conditions. The modularity of the redox-flow batteries put great expectative in its use, nevertheless their investment costs make them few competitive at present. An energy storage system design methodology had been developed along the present chapter. The main goal of this design methodology was to choose an energy storage system able to integrate as much renewable generation as possible in the power system, taking into account the trade off between economic benefits and its capital costs. The first phase of the energy storage system design procedure consists of the evaluation of the economic benefit of the lower thermal power plants generation that results from the operation of several power and energy capacity storage combinations. During the second phase, the capital costs of the several power and energy capacity of the storage systems are computed and the best solutions are identified. In the case study presented for the Portuguese system it is showed that the implementation of the energy storage equipment in power systems with high penetration of renewable power sources, improves the power demand and generation balance and reduces the need for renewable generation curtailment. Some power systems with large amount of hydro installed capacity, such as the Portuguese system, comprise reversible hydro power plants with energy storage capacity, usually used for the energy management improvement. Therefore, two opposite scenarios were considered: the first scenario is related to the pumping hydro system operation accordingly to the historic data; in the second scenario it was assumed that all of the pumping hydro installed capacity was available for helping the integration of the renewable power generation. As far the energy storage systems implementation is concerned, those scenarios allowed the identification of their maximum and minimum capacity requirements. The availability of the installed pumping hydro systems, to cooperate in the integration of the renewable generation appears as a very positive solution for the power unbalance mitigation, especially when considering the costs of completely new energy storage systems implementation. The energy storage system design process needs then to consider the physical constraints of power grid. Therefore, in order to validate the results from that process, namely the power flows imposed by the unit commitment problem outputs, a power system simulator was used (PSS/E - Siemens) for the simulation of the Portuguese transmission network model. All the solutions were validated using the centralized energy storage concept. The solution that does not consider the pumping hydro availability had been validated also with the distributed energy storage at demand and the distributed energy storage at generation concepts. Regarding to the power grid losses, in this specific case study, the three different energy storage locations presented very similar results. 8. References Agbossou, K.; Kolhe, M.; Hamelin, J. & Bose, T. K. (2004). Performance of a Stand-Alone Renewable Energy System Based on Energy Storage as Hydrogen. IEEE Transactions on Energy Conversion, Vol. 19, No. 3, (Sept. 2004), page numbers (633-640), ISSN 0885-8969 Barton, J. P. & Infield, D. G. (2004). Energy Storage and Its Use with Intermittent Renewable Energy. IEEE Transactions on Energy Conversion, Vol. 19, No. 2, (June 2004), page numbers (441-448), ISSN 0885-8969 Chacra, F.; Bastard, P.; Fleury, G. & Clavreul, R. (2005). Impact of Energy Storage Costs on Economical Performance in a Distribution Substation. IEEE Transactions on Power Systems, Vol. 20, No. 2, (May 2005), page numbers (684-691), ISSN 0885-8950 EPRI (2007). Vanadium Redox Flow Batteries: An In-Depth Analysis. EPRI, Palo Alto, CA: 2007. 1014836 ESA (2009). Technology Comparisons, Electricity Storage Association (Online), Available: http://www.electricitystorage.org Faias, S.; Sousa, J.; Castro, R. (2007). Contribution of energy storage systems for power generation and demand balancing with increasing integration of renewable sources: application to the Portuguese power system, Proceedings of the 12th European Conference on Power Electronics and Applications, EPE 2007, pp. 1-10, Vol. 1, ISBN 978-92-75815-10-8, Aalborg, September 2007 Faias, S.; Santos, P.; Matos, F.; Sousa, J.; Castro, R. (2008). Evaluation of energy storage devices for renewable energies integration: Application to a Portuguese wind farm, Proceedings of the 5th International Conference on the European Electricity Market, EEM08, pp. 1-7, Vol. 1, ISBN 978-1-4244-1743-8, Lisbon, May 2008 Gollmer, R.; Nowak, M.; Römisch, W. & Schultz, R. (2000). Unit commitment in power generation – a basic model and some extensions. Annals of Operations Research, Vol. 96, No. 1-4, (November 2000) page numbers (167-189), ISSN 0254-5330 Hammons, T. J. (2006). Integrating Renewable Energy Sources into European Grids, Proceedings of the 41st International Universities Power Engineering Conference, pp. 142-151, Vol. 1, ISBN 978-186135-342-9, Newcastle upon Tyne, September 2006 Hamsic, N.; Schmelter, A.; Mohd, A.; Ortjohann, E.; Schultze, E.; Tuckey, A. & Zimmermann, J. (2007). Increasing Renewable Energy Penetration in Isolated Grids Using a Flywheel Energy Storage System, Proceedings of Powereng 2007 - International Conference on Power Engineering, Energy and Electrical Drives, pp. 195-200, ISBN 978-1-4244-0895-5, Setúbal – Portugal, April 2007 Hoogers, G. (2003). Fuel Cell Technology Handbook, CRC Press, ISBN 0-84-930877-1 Hsu, C. & Lee, W. (1992). Superconducting Magnetic Energy Storage for Power System Applications. IEEE Transactions on Industrial Applications, Vol. 29, No. 5, (Sept.-Oct. 1993), page numbers (990-996), ISSN 0093-9994 Hunt, G. L. (1998). The Great Battery Search. Spectrum IEEE, Vol. 35, No. 11, (Nov. 1998), page numbers (21-28), ISSN 0018-9235 IEA (2005). Variability of Wind Power and other Renewables: Management Options and Strategies. International Energy Agency Publications, Paris Lazarewicz, M. & Rojas, A. (2004). Grid Frequency Regulation by Recycling Electrical Energy in Flywheels, Proceedings of IEEE-Power Engineering Society General Meeting, Vol.2, pp. 2038- 2042, ISBN 0-7803-8465-2, Denver, CO, June 2004 Embedded Energy Storage Systems in the Power Grid for Renewable Energy Sources Integration 87Renewable Energy88 Lerch, E. (2007). Storage of Fluctuating Wind Energy. Proceedings of the 12th European Conference on Power Electronics and Applications, EPE 2007, pp. 1-8, ISBN 978-92-75815-10-8, Aalborg, Denmark, September 2007 Mito, T.; Kawagoe, A.; Chikaraishi, H.; Okumura, K.; Abe, R.; Baba, T.; Yamauchi, K.; Yokota, M.; Henmi, T.; Seo, K.; Hayashi, K.; Iwakuma, M. & Sumiyoshi, F. (2004). Development of UPS-SMES as a Protection From Momentary Voltage Drop. IEEE Transactions on Applied Superconductivity, Vol. 14, No. 2, (June 2004) page numbers (721-726), ISSN 1051-8223 Nomura, S.; Tsutsui, H.; Tsuji-Iio, S. & Shimada, R. (2006). Flexible Power Interconnection With SMES. IEEE Transactions on Applied Superconductivity, Vol. 16, No. 2, (June 2006) page numbers (616-619), ISSN 1051-8223 Ogden, J. M. (1999). Prospects for Building a Hydrogen Infrastructure. Annual Review of Energy and the Environment, Vol. 24, No. 1, page numbers (227-279), ISSN 1056-3466 Ortega-Vazquez, M.; Kirschen, D. (2009). Estimating the Spinning Reserve Requirments in Systems with Significant Wind Power Generation Penetration. IEEE Transactions on Power Systems, Vol. 24, No. 1, (Feruary 2009), page numbers (114-124), ISSN 0885-8950 Ponce de León, C.; Ferrer, A.; García, J. & Walsh, F. (2006). Redox Flow Cells for Energy Conversion. Journal of Power Sources, Vol. 160, No. 1, (Sep. 2006), page numbers (716-732), ISSN 0378-7753 Price, A.; Bartley, S.; Male, S. & Cooley, G. (1999). A Novel Approach to Utility Scale Energy Storage. Power Engineering Journal, Vol. 13, No. 3, (June 1999) page numbers (122-129), ISSN 0950-3366 REN (2008). Electricity - Real Time Information: Technical Data, REN (Online), Available: http://www.centrodeinformacao.ren.pt/EN Riggs, J.; Bedworth, D. & Randhawa, S. (1998). Engineering Economics – Fourth Edition, McGraw-Hill International Publications, ISBN 0-07-115526-0 Rufer, A.; Hotellier, D. & Barrade, P. (2004). A Supercapacitor-Based Energy Storage Substation for Voltage Compensation in Weak Transportation Networks. IEEE Transactions on Power Delivery, Vol. 19, No. 2, (April 2004) page numbers (629-636), ISSN 0885-8977 Schoenung, S.; Eyer, J.; Iannucci, J. & Horgan, S. (1996). Energy Storage for a Competitive Power Market. Annual Review of Energy and the Environment, Vol. 21, No. 1, (November 1996) page numbers (347-370), ISSN 1056-3466 Tande, J. O. G. (2003). Grid Integration of Wind Farms. Wind Energy, Vol. 6, No. 3, (July/September 2003) page numbers (281-295), ISSN 1095-4244 Wang, C. & Shahidehpour, S. M. (1993). Effects of Ramp-Rate Limits on Unit Commitment and Economic Dispatch. IEEE Transactions on Power Systems, Vol. 8, No. 3, (August 1993) page numbers (1341-1350), ISSN 0885-8950 Zhai, N.; Yao, Y.; Zhang, D. & Xu, D. (2006). Design and Optimization for a Supercapacitor Application System, Proceedings of International Conference on Power System Technology, pp. 1-4, ISBN 1-4244-0110-0, Chongqing, October 2006 Single-Phase Grid Connected Converters for Photovoltaic Plants 89Single-Phase Grid Connected Converters for Photovoltaic PlantsEmilio Lorenzani, Giovanni Franceschini, Alberto Bellini and Carla TassoniX Single-Phase Grid Connected Converters for Photovoltaic Plants Emilio Lorenzani1, Giovanni Franceschini1, Alberto Bellini2 and Carla Tassoni1 1University of Parma, Italy 2University of Modena and Reggio Emilia, Italy 1. Introduction Among renewable energy sources photovoltaic systems are one of the most up to date solutions. However their diffusion is limited by relatively high cost in comparison with traditional energy sources. The downward tendency in the price of the PV (Photovoltaic) modules, together with their increasing efficiency, put solid-state inverters under the spot lights as the enabling technology for integrating PV systems into the grid. This chapter will analyze the single-phase grid connected converter usually used in PV domestic rooftop applications. There are two mandatory tasks in grid connected PV systems: the maximization of the energy extracted from the PV panels (1) and the use of a high efficiency topology for the power converter able to inject only active current into the grid (2), i.e. a pure sinusoidal current in phase with the grid voltage. The first task will not be discussed in depth in this chapter and will be mentioned in section 2. The second task is accomplished by using a PLL (Phase Looked Loop) for the grid synchronisation (Chung, 2000), (Arruda et al., 2001), (Silva et al., 2004) and by designing a suitable current controller. In three-phase converters current controllers in a rotating reference frame (d-q frame) are used that provide zero steady-state error and superior disturbance rejection. In single-phase converters achieving zero steady state error at grid frequency is not a simple task. As known, the poor performances of the integral action at frequencies different from zero leads to steady state error and to poor disturbance rejection, making the PI controller not suitable to track a sinusoidal current set point. Interesting solutions were presented in (Yuan et al., 2002), (Teodorescu et al., 2004). In these papers the authors develop the P+resonant (PR) controller for reference tracking in the stationary frame. The use of control system in synchronous reference frame (d-q control), also for single-phase converters, was proposed for stand alone (Ryan & Lorenz, 1997), (Miranda et al., 2005) and for grid connected operations (Cacciato et al., 2008), (Bellini et al., 2008). In section 4 the d-q control is presented and the experimental results will show its nice performances. Section 3 presents the topologies of PV systems, past and present technology, and the possible evolutions oriented towards the increase of the energy extracted by each PV module. 6Renewable Energy88 Lerch, E. (2007). Storage of Fluctuating Wind Energy. Proceedings of the 12th European Conference on Power Electronics and Applications, EPE 2007, pp. 1-8, ISBN 978-92-75815-10-8, Aalborg, Denmark, September 2007 Mito, T.; Kawagoe, A.; Chikaraishi, H.; Okumura, K.; Abe, R.; Baba, T.; Yamauchi, K.; Yokota, M.; Henmi, T.; Seo, K.; Hayashi, K.; Iwakuma, M. & Sumiyoshi, F. (2004). Development of UPS-SMES as a Protection From Momentary Voltage Drop. IEEE Transactions on Applied Superconductivity, Vol. 14, No. 2, (June 2004) page numbers (721-726), ISSN 1051-8223 Nomura, S.; Tsutsui, H.; Tsuji-Iio, S. & Shimada, R. (2006). Flexible Power Interconnection With SMES. IEEE Transactions on Applied Superconductivity, Vol. 16, No. 2, (June 2006) page numbers (616-619), ISSN 1051-8223 Ogden, J. M. (1999). Prospects for Building a Hydrogen Infrastructure. Annual Review of Energy and the Environment, Vol. 24, No. 1, page numbers (227-279), ISSN 1056-3466 Ortega-Vazquez, M.; Kirschen, D. (2009). Estimating the Spinning Reserve Requirments in Systems with Significant Wind Power Generation Penetration. IEEE Transactions on Power Systems, Vol. 24, No. 1, (Feruary 2009), page numbers (114-124), ISSN 0885-8950 Ponce de León, C.; Ferrer, A.; García, J. & Walsh, F. (2006). Redox Flow Cells for Energy Conversion. Journal of Power Sources, Vol. 160, No. 1, (Sep. 2006), page numbers (716-732), ISSN 0378-7753 Price, A.; Bartley, S.; Male, S. & Cooley, G. (1999). A Novel Approach to Utility Scale Energy Storage. Power Engineering Journal, Vol. 13, No. 3, (June 1999) page numbers (122-129), ISSN 0950-3366 REN (2008). Electricity - Real Time Information: Technical Data, REN (Online), Available: http://www.centrodeinformacao.ren.pt/EN Riggs, J.; Bedworth, D. & Randhawa, S. (1998). Engineering Economics – Fourth Edition, McGraw-Hill International Publications, ISBN 0-07-115526-0 Rufer, A.; Hotellier, D. & Barrade, P. (2004). A Supercapacitor-Based Energy Storage Substation for Voltage Compensation in Weak Transportation Networks. IEEE Transactions on Power Delivery, Vol. 19, No. 2, (April 2004) page numbers (629-636), ISSN 0885-8977 Schoenung, S.; Eyer, J.; Iannucci, J. & Horgan, S. (1996). Energy Storage for a Competitive Power Market. Annual Review of Energy and the Environment, Vol. 21, No. 1, (November 1996) page numbers (347-370), ISSN 1056-3466 Tande, J. O. G. (2003). Grid Integration of Wind Farms. Wind Energy, Vol. 6, No. 3, (July/September 2003) page numbers (281-295), ISSN 1095-4244 Wang, C. & Shahidehpour, S. M. (1993). Effects of Ramp-Rate Limits on Unit Commitment and Economic Dispatch. IEEE Transactions on Power Systems, Vol. 8, No. 3, (August 1993) page numbers (1341-1350), ISSN 0885-8950 Zhai, N.; Yao, Y.; Zhang, D. & Xu, D. (2006). Design and Optimization for a Supercapacitor Application System, Proceedings of International Conference on Power System Technology, pp. 1-4, ISBN 1-4244-0110-0, Chongqing, October 2006 Single-Phase Grid Connected Converters for Photovoltaic Plants 89Single-Phase Grid Connected Converters for Photovoltaic PlantsEmilio Lorenzani, Giovanni Franceschini, Alberto Bellini and Carla TassoniX Single-Phase Grid Connected Converters for Photovoltaic Plants Emilio Lorenzani1, Giovanni Franceschini1, Alberto Bellini2 and Carla Tassoni1 1University of Parma, Italy 2University of Modena and Reggio Emilia, Italy 1. Introduction Among renewable energy sources photovoltaic systems are one of the most up to date solutions. However their diffusion is limited by relatively high cost in comparison with traditional energy sources. The downward tendency in the price of the PV (Photovoltaic) modules, together with their increasing efficiency, put solid-state inverters under the spot lights as the enabling technology for integrating PV systems into the grid. This chapter will analyze the single-phase grid connected converter usually used in PV domestic rooftop applications. There are two mandatory tasks in grid connected PV systems: the maximization of the energy extracted from the PV panels (1) and the use of a high efficiency topology for the power converter able to inject only active current into the grid (2), i.e. a pure sinusoidal current in phase with the grid voltage. The first task will not be discussed in depth in this chapter and will be mentioned in section 2. The second task is accomplished by using a PLL (Phase Looked Loop) for the grid synchronisation (Chung, 2000), (Arruda et al., 2001), (Silva et al., 2004) and by designing a suitable current controller. In three-phase converters current controllers in a rotating reference frame (d-q frame) are used that provide zero steady-state error and superior disturbance rejection. In single-phase converters achieving zero steady state error at grid frequency is not a simple task. As known, the poor performances of the integral action at frequencies different from zero leads to steady state error and to poor disturbance rejection, making the PI controller not suitable to track a sinusoidal current set point. Interesting solutions were presented in (Yuan et al., 2002), (Teodorescu et al., 2004). In these papers the authors develop the P+resonant (PR) controller for reference tracking in the stationary frame. The use of control system in synchronous reference frame (d-q control), also for single-phase converters, was proposed for stand alone (Ryan & Lorenz, 1997), (Miranda et al., 2005) and for grid connected operations (Cacciato et al., 2008), (Bellini et al., 2008). In section 4 the d-q control is presented and the experimental results will show its nice performances. Section 3 presents the topologies of PV systems, past and present technology, and the possible evolutions oriented towards the increase of the energy extracted by each PV module. 6Renewable Energy90 A very common topology for single-phase converters is the full-bridge with a line frequency transformer on the AC-side for galvanic isolation purposes. In this configuration the current injected into the grid is the transformer output current. Adopting this solution the DC current component at the transformer input is not directly controlled, and core saturation may appear producing current distortion. A simple solution to avoid core saturation issues is to control the current at transformer input instead of the injected grid current (Ciobotaru et al., 2005). However, this solution leads to lower power factor because of the presence of the transformer magnetizing current, especially at light load conditions. On the other hand, the direct control of the grid current requires a solution to saturation problems without introducing input transformer current measurement. In section 5 an original solution based only on injected current measurement is presented (Bellini et al. 2008). Experimental results show the effectiveness of the proposed method. In new converter designs the tendency is to abandon line frequency transformers because of size, weight and price in favour of high frequency transformers. The presence of high frequency transformer requires several power stages and, as a consequence, increasing efficiency and reducing cost may be a hard task. In low power PV applications it is possible to remove the transformer at all in order to reduce losses, costs and size. The latter configurations are usually referred to as transformerless power converters. Without transformer coupling galvanic isolation between the grid and the DC source is lost and the big parasitic capacitance of the PV panel (Calais & Agelidis, 1998), (Calais et al, 2001) may introduce ground leakage currents. Ground leakage currents increase conducted and radiated electromagnetic emissions, harmonics injected in the grid, losses and the converter has to be disconnected from the grid if the ground leakage currents exceed prefixed limits. Section 6 reports the root causes of the ground leakage currents and some solutions. The best way to reduce them is the use of a suitable converter topology. Inside this section an original full bridge topology with unipolar PWM is presented that ensures a strong reduction of the leakage ground currents. Simulation results are reported that show the effectiveness of the proposed topology. 2. Maximum power exploitation of PV cells The power generation from PV modules suffers from two main shortcomings: the efficiency of electric power generation is very low, especially under low radiation states (1), and the amount of electric power generated by solar arrays is always changing with weather conditions (2). Figure 1 shows the typical variations of the current voltage characteristic of a PV module respectively for different values of irradiance and working temperature. Figure 2 shows the equivalent circuit of a PV cell. A maximum power point tracking (MPPT) algorithm must operate in real-time in PV generation systems. The ideal MPPT algorithm sets the operating point of the PV panel, usually forcing the panel voltage at the Maximum Power Point (MPP). Figure 3 shows an example of the power-voltage characteristic of a photovoltaic field, composed by PV modules connected in series and in parallel for different values of irradiance and partial shading conditions. Most of the MPPT algorithms are based on perturb and observe and on incremental conductance methods. Many improvements to these methods have been presented in order to improve the dynamic performances and to reduce undesired oscillations around the MPP, (Femia et al., 2005), (Fangrui et al, 2008). In high power PV systems the ripple correlation control method can be used that is based on measuring and processing the current and/or voltage ripple created by the switching stage of the power converter (Esram et al., 2006). (a) (b) Fig. 1. Electric characteristics of PV panels (Sanyo Photovoltaic HIP-230HDE1) for different values of irradiance and working temperature. Fig. 2. Equivalent circuit of a PV cell. Fig. 3. Example of power-voltage characteristic of a photovoltaic field for different weather conditions. In order to harvest the maximum energy the ripple at the terminals of the PV modules, must be minimized around the MPP. In single-phase converters since the instantaneous power does not equal the average power, there is an intrinsic fluctuation of the PV module output power at twice the frequency of the grid. The reduction of the output ripple of the PV modules can be obtained only with a power decoupling device that is normally realized by means of large DC capacitors which decrease the lifetime and reliability of the whole system. The DC capacitors are either placed in parallel to the PV modules or to the dc link after a DC-DC converter used for voltage amplification. In a three-phase system the instantaneous output power is constant and no DC large capacitors are needed leading to smaller cost and higher reliability with respect to single-phase inverters. The use of three-phase system also in low power applications could be an interesting solution for newer designs. Single-Phase Grid Connected Converters for Photovoltaic Plants 91Renewable Energy90 A very common topology for single-phase converters is the full-bridge with a line frequency transformer on the AC-side for galvanic isolation purposes. In this configuration the current injected into the grid is the transformer output current. Adopting this solution the DC current component at the transformer input is not directly controlled, and core saturation may appear producing current distortion. A simple solution to avoid core saturation issues is to control the current at transformer input instead of the injected grid current (Ciobotaru et al., 2005). However, this solution leads to lower power factor because of the presence of the transformer magnetizing current, especially at light load conditions. On the other hand, the direct control of the grid current requires a solution to saturation problems without introducing input transformer current measurement. In section 5 an original solution based only on injected current measurement is presented (Bellini et al. 2008). Experimental results show the effectiveness of the proposed method. In new converter designs the tendency is to abandon line frequency transformers because of size, weight and price in favour of high frequency transformers. The presence of high frequency transformer requires several power stages and, as a consequence, increasing efficiency and reducing cost may be a hard task. In low power PV applications it is possible to remove the transformer at all in order to reduce losses, costs and size. The latter configurations are usually referred to as transformerless power converters. Without transformer coupling galvanic isolation between the grid and the DC source is lost and the big parasitic capacitance of the PV panel (Calais & Agelidis, 1998), (Calais et al, 2001) may introduce ground leakage currents. Ground leakage currents increase conducted and radiated electromagnetic emissions, harmonics injected in the grid, losses and the converter has to be disconnected from the grid if the ground leakage currents exceed prefixed limits. Section 6 reports the root causes of the ground leakage currents and some solutions. The best way to reduce them is the use of a suitable converter topology. Inside this section an original full bridge topology with unipolar PWM is presented that ensures a strong reduction of the leakage ground currents. Simulation results are reported that show the effectiveness of the proposed topology. 2. Maximum power exploitation of PV cells The power generation from PV modules suffers from two main shortcomings: the efficiency of electric power generation is very low, especially under low radiation states (1), and the amount of electric power generated by solar arrays is always changing with weather conditions (2). Figure 1 shows the typical variations of the current voltage characteristic of a PV module respectively for different values of irradiance and working temperature. Figure 2 shows the equivalent circuit of a PV cell. A maximum power point tracking (MPPT) algorithm must operate in real-time in PV generation systems. The ideal MPPT algorithm sets the operating point of the PV panel, usually forcing the panel voltage at the Maximum Power Point (MPP). Figure 3 shows an example of the power-voltage characteristic of a photovoltaic field, composed by PV modules connected in series and in parallel for different values of irradiance and partial shading conditions. Most of the MPPT algorithms are based on perturb and observe and on incremental conductance methods. Many improvements to these methods have been presented in order to improve the dynamic performances and to reduce undesired oscillations around the MPP, (Femia et al., 2005), (Fangrui et al, 2008). In high power PV systems the ripple correlation control method can be used that is based on measuring and processing the current and/or voltage ripple created by the switching stage of the power converter (Esram et al., 2006). (a) (b) Fig. 1. Electric characteristics of PV panels (Sanyo Photovoltaic HIP-230HDE1) for different values of irradiance and working temperature. Fig. 2. Equivalent circuit of a PV cell. Fig. 3. Example of power-voltage characteristic of a photovoltaic field for different weather conditions. In order to harvest the maximum energy the ripple at the terminals of the PV modules, must be minimized around the MPP. In single-phase converters since the instantaneous power does not equal the average power, there is an intrinsic fluctuation of the PV module output power at twice the frequency of the grid. The reduction of the output ripple of the PV modules can be obtained only with a power decoupling device that is normally realized by means of large DC capacitors which decrease the lifetime and reliability of the whole system. The DC capacitors are either placed in parallel to the PV modules or to the dc link after a DC-DC converter used for voltage amplification. In a three-phase system the instantaneous output power is constant and no DC large capacitors are needed leading to smaller cost and higher reliability with respect to single-phase inverters. The use of three-phase system also in low power applications could be an interesting solution for newer designs. Single-Phase Grid Connected Converters for Photovoltaic Plants 91Renewable Energy92 3. Topologies for grid connected photovoltaic systems The basic elements of a PV system are the modules that are usually series-connected. A series of PV modules is usually called a PV string. If the voltage of the PV string is always higher than the peak voltage of the grid the PV converter does not require a step-up stage. In this case higher efficiency can be obtained because a single stage full-bridge converter can be used. Otherwise, a DC-DC converter or a transformer must be added for voltage amplification reducing efficiency. A PV system is the combination of PV fields and the related power converters. The peak current that can be delivered by one string is determined by the PV module characteristics, figure 1. To achieve a higher power level several strings can be connected in parallel as shown in figure 4a. In this way, a single converter can be used reducing the cost and the losses of the static energy conversion. In this topology, usually referred to as Central Converter, the lack of individual MPPT for each string does not permit to harvest the maximum electric power from PV modules, especially when shading or different orientation of modules occurs. This major shortcoming results in avoiding this simple topology in newer photovoltaic system designs. (a) Central Converter (b) String converter (c) Multi-string converter (d) AC module Fig. 4. Configurations of the power converter for a grid connected PV System, central converter (a), string converter (b), multi-string converter (c) and AC module (d). Other options are possible as sketched in Figure 4. The string converter topology is shown in figure 4(b). This configuration does not employ the parallel connections of the strings. Each string has its own MPPT and is completely independent from each other. Therefore it is easy to build PV systems with different orientations, shading conditions and number of PV modules for each string. A disadvantage of string-converters in comparison to central converters is the higher price per kW. String converters are often build only as single-phase converters due to the low power level. A very common classic topology is the full-bridge with a line frequency transformer on the AC-side for galvanic isolation and for voltage step-up. This architecture is analyzed in section 5. The multi-string converter (figure 4(c)) manages two or three strings, and provides independent MPPT by different DC-DC converters. In this case a two-stage configuration is mandatory. Thanks to these additional DC-DC stages, used also in some string converter, it is possible to obtain a very wide input voltage range which gives to the user big freedom in designing of the photovoltaic field. In these topologies the modules of one string have to be well matched and should be installed in the same orientation to achieve a high energy harvest. The photovoltaic energy harvesting can be maximized by using an individual MPPT for each PV module. The low power level permits to integrate the converter into the housing of the PV module, that is called AC module (figure 4(d)). The AC module is connected directly to the grid and no DC wirings are needed between PV modules. Despite their simple use and installation the low power level of AC modules leads to higher cost per watt. The major issue of this solution is the lifetime of the actual converters that is smaller than the lifetime of a PV module (20 years and more). When it will be comparable this solution will become interesting. 4. Current Controllers for Single-Phase Grid Connected Converters 4.1 d-q current controller The DC-AC stage of a PV system must inject active grid current only, i.e. a pure sinusoidal current in phase with the grid voltage. To satisfy this condition the steady state error between the desired grid current and the actual one must be tightly zero at grid frequency. To this aim many control architectures have been presented in literature to overcome the drawbacks of the simple PI controller. As known, the poor performances of the integral action at frequency different from zero lead to steady state error and to poor disturbance rejection making the PI controller not suitable to track a sinusoidal set point in a stationary reference frame. The performances of PI controllers can be improved adding a feed-forward compensation to its output. The feed-forward value is computed in order to obtain at the output of the power converter the same voltage imposed by the grid in case of a zero contribution of the PI controller. The PI + feed-forward controller is used in some experimental results shown in section 6. An interesting alternative is the P+resonant (PR) controller that introduces an infinite gain at a selected resonant frequency providing a theoretical zero steady-state error at that frequency. Actually the gain at resonant frequency is limited by stability problems. A further option is the use of a control system in a reference frame synchronous with the grid frequency (d-q reference frame). The d-q control allows an infinite control loop gain at grid frequency and a superior disturbance rejection. Hence, it can increase efficiency cancelling reactive current delivered to the grid. Mimicking the technique used in case of the single-phase transport delay PLL (Arruda et al, 2001), (Silva et al., 2004) to obtain the d-q current components in synchronous reference a current i t  I sin t  i , orthogonal to the grid current i t  I cos t i , is introduced. Appling the Park transformation Id and Iq can be easily computed. Single-Phase Grid Connected Converters for Photovoltaic Plants 93Renewable Energy92 3. Topologies for grid connected photovoltaic systems The basic elements of a PV system are the modules that are usually series-connected. A series of PV modules is usually called a PV string. If the voltage of the PV string is always higher than the peak voltage of the grid the PV converter does not require a step-up stage. In this case higher efficiency can be obtained because a single stage full-bridge converter can be used. Otherwise, a DC-DC converter or a transformer must be added for voltage amplification reducing efficiency. A PV system is the combination of PV fields and the related power converters. The peak current that can be delivered by one string is determined by the PV module characteristics, figure 1. To achieve a higher power level several strings can be connected in parallel as shown in figure 4a. In this way, a single converter can be used reducing the cost and the losses of the static energy conversion. In this topology, usually referred to as Central Converter, the lack of individual MPPT for each string does not permit to harvest the maximum electric power from PV modules, especially when shading or different orientation of modules occurs. This major shortcoming results in avoiding this simple topology in newer photovoltaic system designs. (a) Central Converter (b) String converter (c) Multi-string converter (d) AC module Fig. 4. Configurations of the power converter for a grid connected PV System, central converter (a), string converter (b), multi-string converter (c) and AC module (d). Other options are possible as sketched in Figure 4. The string converter topology is shown in figure 4(b). This configuration does not employ the parallel connections of the strings. Each string has its own MPPT and is completely independent from each other. Therefore it is easy to build PV systems with different orientations, shading conditions and number of PV modules for each string. A disadvantage of string-converters in comparison to central converters is the higher price per kW. String converters are often build only as single-phase converters due to the low power level. A very common classic topology is the full-bridge with a line frequency transformer on the AC-side for galvanic isolation and for voltage step-up. This architecture is analyzed in section 5. The multi-string converter (figure 4(c)) manages two or three strings, and provides independent MPPT by different DC-DC converters. In this case a two-stage configuration is mandatory. Thanks to these additional DC-DC stages, used also in some string converter, it is possible to obtain a very wide input voltage range which gives to the user big freedom in designing of the photovoltaic field. In these topologies the modules of one string have to be well matched and should be installed in the same orientation to achieve a high energy harvest. The photovoltaic energy harvesting can be maximized by using an individual MPPT for each PV module. The low power level permits to integrate the converter into the housing of the PV module, that is called AC module (figure 4(d)). The AC module is connected directly to the grid and no DC wirings are needed between PV modules. Despite their simple use and installation the low power level of AC modules leads to higher cost per watt. The major issue of this solution is the lifetime of the actual converters that is smaller than the lifetime of a PV module (20 years and more). When it will be comparable this solution will become interesting. 4. Current Controllers for Single-Phase Grid Connected Converters 4.1 d-q current controller The DC-AC stage of a PV system must inject active grid current only, i.e. a pure sinusoidal current in phase with the grid voltage. To satisfy this condition the steady state error between the desired grid current and the actual one must be tightly zero at grid frequency. To this aim many control architectures have been presented in literature to overcome the drawbacks of the simple PI controller. As known, the poor performances of the integral action at frequency different from zero lead to steady state error and to poor disturbance rejection making the PI controller not suitable to track a sinusoidal set point in a stationary reference frame. The performances of PI controllers can be improved adding a feed-forward compensation to its output. The feed-forward value is computed in order to obtain at the output of the power converter the same voltage imposed by the grid in case of a zero contribution of the PI controller. The PI + feed-forward controller is used in some experimental results shown in section 6. An interesting alternative is the P+resonant (PR) controller that introduces an infinite gain at a selected resonant frequency providing a theoretical zero steady-state error at that frequency. Actually the gain at resonant frequency is limited by stability problems. A further option is the use of a control system in a reference frame synchronous with the grid frequency (d-q reference frame). The d-q control allows an infinite control loop gain at grid frequency and a superior disturbance rejection. Hence, it can increase efficiency cancelling reactive current delivered to the grid. Mimicking the technique used in case of the single-phase transport delay PLL (Arruda et al, 2001), (Silva et al., 2004) to obtain the d-q current components in synchronous reference a current i t  I sin t  i , orthogonal to the grid current i t  I cos t i , is introduced. Appling the Park transformation Id and Iq can be easily computed. Single-Phase Grid Connected Converters for Photovoltaic Plants 93Renewable Energy94 Let v t  V cos t v  and i t  I cos t i  be the grid voltage and the grid current, where  is the grid pulsation, v and i are respectively the voltage and the current displacement, it holds: IdIq cos t  sin t -sin t  cos t ii (1) where the angle  t  t v is obtained once the PLL is locked to the grid voltage angle. Matrix computations lead to the following results: Id  I cos v  i  I cosIq  I sin v  i  I sin (2) Hence, Id and –Iq are respectively the amplitude of the active grid current, in phase with the grid voltage, and the amplitude of the reactive grid current, in quadrature with the grid voltage; cos is the power factor. Being the system a single-phase one, only one manipulated variable can be managed. This variable, named in Figure 5 Vβ, depends, generally speaking, on park transformation reference choice, and drives the single-phase DC/AC converter in order to obtain the desired grid current. Vβ is a output of the Park inverse transformation, whose inputs are: 1) the Vd voltage, i.e. the output of the active current (Id) control loop, 2) the Vq voltage, i.e. the output of reactive current (-Iq) control loop. The set-point of the reactive current control loop is set to zero because in ideal conditions only active current must be supplied. A simpler choice would require only one current loop, fixing Vq=0. However, if the Iq control loop is opened, the presence of reactive voltage drops and system non ideal behavior will introduce a reactive current component and then a wrong current angle, as it will be shown in the following by simulation. The right angle is achieved only if the reactive current loop is closed with a zero set-point. In fact, in ideal conditions, only the Id current component must appear and the actual Iq current represents a measure of the system error. Fig. 5. Block diagram of d-q control structure. 4.2 Simulation Results Numerical simulations were used to assess the performances of the proposed d-q control structure. The MATLAB simulink environment was used to model the whole system, where the power converter is modeled by PLECS toolbox. A fixed step time of 5 ms was used and all simulations refer to a step set-point of Id with an amplitude of 10A. At first some simulations were made to verify the system performances in case of Iq current loop opened and Vq set to zero. In this case, the presence of the Id control loop will fix the active grid current at the right value. However, because of reactive voltage drops and system non ideal behavior a reactive current component will appear and the grid current amplitude will increase beyond the desired level. Specifically, the synchronous reference frame control forces a zero steady state error in d-axis but this result is reached increasing the grid current amplitude instead of adjusting its angle. Figure 6 shows the grid voltage and current under these conditions: the amplitude of the grid current is higher than requested because a remarkable amount of reactive current is injected into the grid. In these charts the grid current amplitude was amplified (x10) in order to identify easily the current displacement. The resulting power factor and then system performances are very poor. Then simulations were made relying on the structure of figure 5. Thanks to the action of the Iq axis control loop the reactive current is forced to zero by integral action. Figure 7 shows clearly that only active current is now present. A unitary power factor is obtained and the amplitude of the grid current reaches its set-point (10A). Figure 8 highlights the dynamic response of Id and Iq: in a few periods d-q current control ensures zero steady state error as expected. Fig. 6 Simulation results with simple control structure (Vq=0). Grid voltage (violet trace) and amplified (x10) grid current (red one). Fig. 7 Simulation results with d-q control. Grid voltage (violet trace) and amplified (x10) grid current (red one). Fig. 8 Simulation results with d-q control. Dynamic response of Id (green trace) and Iq (blue one). Single-Phase Grid Connected Converters for Photovoltaic Plants 95Renewable Energy94 Let v t  V cos t v  and i t  I cos t i  be the grid voltage and the grid current, where  is the grid pulsation, v and i are respectively the voltage and the current displacement, it holds: IdIq cos t  sin t -sin t  cos t ii (1) where the angle  t  t v is obtained once the PLL is locked to the grid voltage angle. Matrix computations lead to the following results: Id  I cos v  i  I cosIq  I sin v  i  I sin (2) Hence, Id and –Iq are respectively the amplitude of the active grid current, in phase with the grid voltage, and the amplitude of the reactive grid current, in quadrature with the grid voltage; cos is the power factor. Being the system a single-phase one, only one manipulated variable can be managed. This variable, named in Figure 5 Vβ, depends, generally speaking, on park transformation reference choice, and drives the single-phase DC/AC converter in order to obtain the desired grid current. Vβ is a output of the Park inverse transformation, whose inputs are: 1) the Vd voltage, i.e. the output of the active current (Id) control loop, 2) the Vq voltage, i.e. the output of reactive current (-Iq) control loop. The set-point of the reactive current control loop is set to zero because in ideal conditions only active current must be supplied. A simpler choice would require only one current loop, fixing Vq=0. However, if the Iq control loop is opened, the presence of reactive voltage drops and system non ideal behavior will introduce a reactive current component and then a wrong current angle, as it will be shown in the following by simulation. The right angle is achieved only if the reactive current loop is closed with a zero set-point. In fact, in ideal conditions, only the Id current component must appear and the actual Iq current represents a measure of the system error. Fig. 5. Block diagram of d-q control structure. 4.2 Simulation Results Numerical simulations were used to assess the performances of the proposed d-q control structure. The MATLAB simulink environment was used to model the whole system, where the power converter is modeled by PLECS toolbox. A fixed step time of 5 ms was used and all simulations refer to a step set-point of Id with an amplitude of 10A. At first some simulations were made to verify the system performances in case of Iq current loop opened and Vq set to zero. In this case, the presence of the Id control loop will fix the active grid current at the right value. However, because of reactive voltage drops and system non ideal behavior a reactive current component will appear and the grid current amplitude will increase beyond the desired level. Specifically, the synchronous reference frame control forces a zero steady state error in d-axis but this result is reached increasing the grid current amplitude instead of adjusting its angle. Figure 6 shows the grid voltage and current under these conditions: the amplitude of the grid current is higher than requested because a remarkable amount of reactive current is injected into the grid. In these charts the grid current amplitude was amplified (x10) in order to identify easily the current displacement. The resulting power factor and then system performances are very poor. Then simulations were made relying on the structure of figure 5. Thanks to the action of the Iq axis control loop the reactive current is forced to zero by integral action. Figure 7 shows clearly that only active current is now present. A unitary power factor is obtained and the amplitude of the grid current reaches its set-point (10A). Figure 8 highlights the dynamic response of Id and Iq: in a few periods d-q current control ensures zero steady state error as expected. Fig. 6 Simulation results with simple control structure (Vq=0). Grid voltage (violet trace) and amplified (x10) grid current (red one). Fig. 7 Simulation results with d-q control. Grid voltage (violet trace) and amplified (x10) grid current (red one). Fig. 8 Simulation results with d-q control. Dynamic response of Id (green trace) and Iq (blue one). Single-Phase Grid Connected Converters for Photovoltaic Plants 95Renewable Energy96 4.3 Experimental Results Several experiments were made in order to evaluate the performances of the d-q control. The designed prototype is based on a 56F8323 DSP that implements the control algorithm, the digital filtering, the grid synchronization, the reference frame transformation and the PWM modulation. Specifically a unipolar PWM modulation is adopted with a switching frequency of 10 kHz. The power stage is implemented on a 600 V, 25 A Intelligent Power Module (IPM). Using unipolar modulation a simple LC filter can be used with L = 2 mH, and C = 1.5 F. The detailed design and optimization of the output filter is not investigated in this paper. A 400V low distortion DC voltage source was adopted. Figure 9 and Figure 10 show the performances of the d-q control architecture. Figure 9 refers to a step set-point of Id from 0 to 8A while the set point of Iq is fixed to zero. Under these conditions, the set-point is reached in about 10 ms, i.e. less than a period. The dynamic behavior of the currents is in a nice agreement with simulation results of Figure 8 and, as theoretically predicted, the d-q control keeps the current tightly at its set-point. Figure 10 shows the steady state behaviour of the d-q control for a set-point of Id = 7A. The injected grid current shows only a low distortion content in correspondence of voltage zero crossing. Fig. 9. Experimental Results. Step response: Id (green solid line), Iq (blue solid line) and Id* set-point (red solid line). Fig. 10. Experimental Results. Stationary conditions: grid voltage (red solid line) and injected grid current (blue solid line). 5. Core saturation compensation strategies for PV power converters with line frequency transformers 5.1 Control of injected grid current in line frequency transformer PV converter A very common topology for single-phase converters is the full-bridge voltage source inverter (VSI), driven by unipolar PWM, with a line frequency transformer on the AC-side for galvanic isolation (Figure 11). The line frequency transformer is a safe solution in order to match the electrical safety standards and to block out the DC current component. Moreover, it can be used to adjust the voltage level without requiring a DC/DC power converter. However, the transformer features increased size, weight, and price of the converter. The use of a toroidal transformer can partly reduce these drawbacks. In order to achieve higher efficiency the inverter strategy control must inject only active grid current, i.e. a pure sinusoidal current in phase with the grid voltage. To this aim injected grid current must be controlled directly providing zero steady state error at grid frequency. For a line frequency transformer PV converter, the control of the injected grid current stands for the control of the transformer output current. In this way, the DC current component at transformer input is not directly controlled and core saturation and thus current distortion may appear. A simple solution for avoiding core magnetic saturation, if current measurement is tuned properly, is to control the current at transformer input instead of the grid current. This solution leads to poorer performances at light loads conditions because of the transformer magnetizing current. On the other hand, the direct control of the injected grid current without the measurement of current at transformer input is affected by saturation problems. In (Bellini et al. 2008) the authors presented a solution based only on injected grid current measurement. Fig. 11. Block diagram of a line frequency transformer PV converter. If the control loop is closed at the transformer output a DC voltage and current bias may appear at the transformer input when a non ideal DC-AC converter feeds the transformer. In fact, the transformer blocks out the DC current component, hence the output current control is not able to manage the DC component at the transformer input. This bias can lead to magnetic core saturation and then to harmonic distortion of input and output currents. Figure 12 shows the effects of the DC voltage bias: the magnetic flux, that is, roughly, the integral of the voltage, will saturate asymmetrically. Depending on the sign of the bias, saturation will affect positive or negative semi-periods only. In correspondence of flux saturation the e.m.f. will decrease and the input current will increase accordingly, see Figure 12. Single-Phase Grid Connected Converters for Photovoltaic Plants 97Renewable Energy96 4.3 Experimental Results Several experiments were made in order to evaluate the performances of the d-q control. The designed prototype is based on a 56F8323 DSP that implements the control algorithm, the digital filtering, the grid synchronization, the reference frame transformation and the PWM modulation. Specifically a unipolar PWM modulation is adopted with a switching frequency of 10 kHz. The power stage is implemented on a 600 V, 25 A Intelligent Power Module (IPM). Using unipolar modulation a simple LC filter can be used with L = 2 mH, and C = 1.5 F. The detailed design and optimization of the output filter is not investigated in this paper. A 400V low distortion DC voltage source was adopted. Figure 9 and Figure 10 show the performances of the d-q control architecture. Figure 9 refers to a step set-point of Id from 0 to 8A while the set point of Iq is fixed to zero. Under these conditions, the set-point is reached in about 10 ms, i.e. less than a period. The dynamic behavior of the currents is in a nice agreement with simulation results of Figure 8 and, as theoretically predicted, the d-q control keeps the current tightly at its set-point. Figure 10 shows the steady state behaviour of the d-q control for a set-point of Id = 7A. The injected grid current shows only a low distortion content in correspondence of voltage zero crossing. Fig. 9. Experimental Results. Step response: Id (green solid line), Iq (blue solid line) and Id* set-point (red solid line). Fig. 10. Experimental Results. Stationary conditions: grid voltage (red solid line) and injected grid current (blue solid line). 5. Core saturation compensation strategies for PV power converters with line frequency transformers 5.1 Control of injected grid current in line frequency transformer PV converter A very common topology for single-phase converters is the full-bridge voltage source inverter (VSI), driven by unipolar PWM, with a line frequency transformer on the AC-side for galvanic isolation (Figure 11). The line frequency transformer is a safe solution in order to match the electrical safety standards and to block out the DC current component. Moreover, it can be used to adjust the voltage level without requiring a DC/DC power converter. However, the transformer features increased size, weight, and price of the converter. The use of a toroidal transformer can partly reduce these drawbacks. In order to achieve higher efficiency the inverter strategy control must inject only active grid current, i.e. a pure sinusoidal current in phase with the grid voltage. To this aim injected grid current must be controlled directly providing zero steady state error at grid frequency. For a line frequency transformer PV converter, the control of the injected grid current stands for the control of the transformer output current. In this way, the DC current component at transformer input is not directly controlled and core saturation and thus current distortion may appear. A simple solution for avoiding core magnetic saturation, if current measurement is tuned properly, is to control the current at transformer input instead of the grid current. This solution leads to poorer performances at light loads conditions because of the transformer magnetizing current. On the other hand, the direct control of the injected grid current without the measurement of current at transformer input is affected by saturation problems. In (Bellini et al. 2008) the authors presented a solution based only on injected grid current measurement. Fig. 11. Block diagram of a line frequency transformer PV converter. If the control loop is closed at the transformer output a DC voltage and current bias may appear at the transformer input when a non ideal DC-AC converter feeds the transformer. In fact, the transformer blocks out the DC current component, hence the output current control is not able to manage the DC component at the transformer input. This bias can lead to magnetic core saturation and then to harmonic distortion of input and output currents. Figure 12 shows the effects of the DC voltage bias: the magnetic flux, that is, roughly, the integral of the voltage, will saturate asymmetrically. Depending on the sign of the bias, saturation will affect positive or negative semi-periods only. In correspondence of flux saturation the e.m.f. will decrease and the input current will increase accordingly, see Figure 12. Single-Phase Grid Connected Converters for Photovoltaic Plants 97Renewable Energy98 5.2 Saturation Compensation Strategy In order to avoid core saturation and then current distortion a suitable algorithm was developed to compensate the DC bias effects. At first, the algorithm must identify the semi-period affected by saturation (i.e. the ”sign” of saturation), then it must estimate its entity. Neglecting voltage drop across stray inductance the maximum and the minimum values of magnetic flux are synchronous with the zero crossings of the supply voltage waveform. Hence, it is expected that the effect of saturation will appear around zero crossings of supply voltage. Comparing the output current with its set-point value in correspondence of the zero crossings of supply voltage, two saturation indexes referred to as positive SIP and negative SIN values can be computed: 2*2ZCZCT TP grid gridT TSI i i dt  (3) *ZCZCT TN grid gridT TSI i i dt   (4) where i grid and i*grid are the actual grid current and the desired grid current waveform. From a theoretical point of view the time interval ΔTZC must be less then or, at least, equal to T/4. A time integration value of T/10 is sufficient to obtain a robust estimation of saturation core conditions. The combination of SIP and SIN is a nice saturation estimator that can lead to an effective DC bias compensation as shown in the following. If a positive DC current component 0I is superimposed to the current at transformer input, saturation affects the positive values of the flux, as reported in Figure 13. Hence the index SIP will be positive while SIN = 0. On the contrary, if a negative DC current component 0I is present, saturation will affect the negative flux values. Hence SIP = 0 and SIN > 0. In case of absence of DC current component ( 0I = 0), saturation is not present and SIP = SIN = 0. Computing and summing the two saturation indexes every period a rough estimation ( 0 _ estI ) of DC current component can be made: 0 _ estI = SIP + SIN. Thanks to the estimation of 0I the effect of the DC current component can be compensated with the simple feedback control detailed in Figure 14, where V* is the output of the current regulator. The control loop forces a zero DC current component 0I thus preventing magnetic saturation. This compensation strategy does not require DC current measurements, thus avoiding intrinsically transducer offset issues. Fig. 12. Effect of a positive DC bias on the current at transformer input. Core flux (green solid line), transformer input (blue solid line) and output (red solid line) currents. Fig. 13. Computation of saturation indexes. Core flux (green trace), current set point (black trace) and transformer output current (red one). Fig. 14. Block diagram of the saturation compensation scheme. 5.3 Experimental Results Several experiments were made in order to evaluate the performances of the saturation compensation strategy. Experiments were made with the prototype described in section 4.1. The saturation compensation strategy is applied to two different current controllers: PI+feed forward and d-q synchronous reference frame controllers. The amplitude of the desired grid current, i.e. the set-point of the current controller, is 8A in both cases. Figure 15a shows the grid voltage, with an intrinsic distortion not correlated to the PV system, the grid current (transformer output current) and the transformer input current in case of PI regulator and feed-forward without saturation compensation. As it can be noted, Single-Phase Grid Connected Converters for Photovoltaic Plants 99Renewable Energy98 5.2 Saturation Compensation Strategy In order to avoid core saturation and then current distortion a suitable algorithm was developed to compensate the DC bias effects. At first, the algorithm must identify the semi-period affected by saturation (i.e. the ”sign” of saturation), then it must estimate its entity. Neglecting voltage drop across stray inductance the maximum and the minimum values of magnetic flux are synchronous with the zero crossings of the supply voltage waveform. Hence, it is expected that the effect of saturation will appear around zero crossings of supply voltage. Comparing the output current with its set-point value in correspondence of the zero crossings of supply voltage, two saturation indexes referred to as positive SIP and negative SIN values can be computed: 2*2ZCZCT TP grid gridT TSI i i dt  (3) *ZCZCT TN grid gridT TSI i i dt   (4) where i grid and i*grid are the actual grid current and the desired grid current waveform. From a theoretical point of view the time interval ΔTZC must be less then or, at least, equal to T/4. A time integration value of T/10 is sufficient to obtain a robust estimation of saturation core conditions. The combination of SIP and SIN is a nice saturation estimator that can lead to an effective DC bias compensation as shown in the following. If a positive DC current component 0I is superimposed to the current at transformer input, saturation affects the positive values of the flux, as reported in Figure 13. Hence the index SIP will be positive while SIN = 0. On the contrary, if a negative DC current component 0I is present, saturation will affect the negative flux values. Hence SIP = 0 and SIN > 0. In case of absence of DC current component ( 0I = 0), saturation is not present and SIP = SIN = 0. Computing and summing the two saturation indexes every period a rough estimation ( 0 _ estI ) of DC current component can be made: 0 _ estI = SIP + SIN. Thanks to the estimation of 0I the effect of the DC current component can be compensated with the simple feedback control detailed in Figure 14, where V* is the output of the current regulator. The control loop forces a zero DC current component 0I thus preventing magnetic saturation. This compensation strategy does not require DC current measurements, thus avoiding intrinsically transducer offset issues. Fig. 12. Effect of a positive DC bias on the current at transformer input. Core flux (green solid line), transformer input (blue solid line) and output (red solid line) currents. Fig. 13. Computation of saturation indexes. Core flux (green trace), current set point (black trace) and transformer output current (red one). Fig. 14. Block diagram of the saturation compensation scheme. 5.3 Experimental Results Several experiments were made in order to evaluate the performances of the saturation compensation strategy. Experiments were made with the prototype described in section 4.1. The saturation compensation strategy is applied to two different current controllers: PI+feed forward and d-q synchronous reference frame controllers. The amplitude of the desired grid current, i.e. the set-point of the current controller, is 8A in both cases. Figure 15a shows the grid voltage, with an intrinsic distortion not correlated to the PV system, the grid current (transformer output current) and the transformer input current in case of PI regulator and feed-forward without saturation compensation. As it can be noted, Single-Phase Grid Connected Converters for Photovoltaic Plants 99Renewable Energy100 current distortion appears in correspondence of the voltage zero crossing at the end of a positive voltage semi-period, hence saturation is due to a positive DC bias at the transformer input. Figure 15b shows the effects of the saturation compensation strategy. (a) PI regulator + feed forward control without saturation compensation (b) PI regulator + feed forward control with saturation compensation (c) d-q control without saturation compensation (d) d-q control with saturation compensation Fig. 15. Experimental Results. Time waveforms of grid voltage (violet solid line), grid current (red solid line) and transformer input current (blue solid line). The same experimental results are made using the d-q control without (Figure 15c) and with (Figure 15d) saturation compensation. As expected, in this case the steady-state error is zero and the better disturbance rejection of the d-q control damps harmonic distortion of the current. The proposed compensation strategy results very effective for both control schemes. 6. Transformerless grid-connected converters In new converter designs the trend is to abandon line frequency transformers because of size, weight and price in favour of high frequency transformers. The presence of high frequency transformer requires several power stages and, as a consequence, increasing efficiency and reducing cost is a hard task. In low power PV applications it is possible to remove transformer at all (transformerless configurations). The absence of the line frequency transformers does not ensure the elimination of DC grid current. The DC component is mainly caused by an offset in the grid current measurement that is usually realized through Hall effect sensors. As known, these sensors are strongly affected by DC offset drift. A current sensor based on fluxgate technology could be adopted to reduce the DC offset drift. The absence of the line frequency transformer also causes two major drawbacks: safety issue in case of isolation faults, and wrong operation because of ground leakage currents. The first issue occurs since the PV generator is not isolated from the grid, as a consequence isolation faults at the PV generator or at DC-wiring can get a safety hazard in case of indirect electric shock. Hence, transformerless inverters are equipped with Residual Current Devices (RCD), which monitor the ground leakage currents during operation. When a dangerous leakage current is recognized, the inverter is immediately disconnected from the AC-grid. A typical configuration for a transformerless power converter, based on H-brige configuration, is reported in Figure 16. Fig. 16. Schematic representation of a transformerless power converter, based on H bridge, for PV systems. Leakage ground currents are shown as dashed lines. The second issue is created by the large parasitic capacitance of the PV panel (10-100nF/kWp) that results in high ground leakage currents in case of variations of common mode voltage at the output of the power converter. Impulsive common mode voltages will result in large ground leakage currents because the power converter is coupled to a resonant circuit with a small damping coefficient. The resonant circuit is composed by the earth capacitance, the converter, the AC filter and the grid. A blocking filter is not easy to tune, since the earth capacitance changes with environmental conditions. The damping of the resonant circuit is small since it is mandatory to reduce dissipative terms and thus to achieve a high efficiency. Therefore the best remedial strategy for ground leakage current is the reduction of the excitation, i.e. of variation of common mode voltage. To this aim it could be necessary to adopt topologies free from common-mode voltage variations (Lopez et al., 2007), (Gonzales et al., 2007). Converter topologies intrinsically safe from leakage currents are the Half-Bridge (HB) and the Neutral Point Clamped (NPC) ones (Gonzales et al., 2008), (Lopez et al., 2006). However, these configurations require twice the input voltage to obtain the same voltage level that is one of the major drawbacks for a PV system. On the other hand, a full-bridge converter driven by bipolar PWM (Gonzales et al 2007) exploits the full voltage rails but reduces common mode voltage with lower efficiency than HB of NPC configurations. From a power quality point of view all these solutions present current ripple at switching frequency. For power converters with transformers the full bridge converter with unipolar PWM is widespread, since it represents a nice trade-off between efficiency, complexity and price. The biggest advantages of unipolar PWM are that the current ripple appears at twice the switching frequency and that the voltage across the AC output filter is unipolar, thus Single-Phase Grid Connected Converters for Photovoltaic Plants 101Renewable Energy100 current distortion appears in correspondence of the voltage zero crossing at the end of a positive voltage semi-period, hence saturation is due to a positive DC bias at the transformer input. Figure 15b shows the effects of the saturation compensation strategy. (a) PI regulator + feed forward control without saturation compensation (b) PI regulator + feed forward control with saturation compensation (c) d-q control without saturation compensation (d) d-q control with saturation compensation Fig. 15. Experimental Results. Time waveforms of grid voltage (violet solid line), grid current (red solid line) and transformer input current (blue solid line). The same experimental results are made using the d-q control without (Figure 15c) and with (Figure 15d) saturation compensation. As expected, in this case the steady-state error is zero and the better disturbance rejection of the d-q control damps harmonic distortion of the current. The proposed compensation strategy results very effective for both control schemes. 6. Transformerless grid-connected converters In new converter designs the trend is to abandon line frequency transformers because of size, weight and price in favour of high frequency transformers. The presence of high frequency transformer requires several power stages and, as a consequence, increasing efficiency and reducing cost is a hard task. In low power PV applications it is possible to remove transformer at all (transformerless configurations). The absence of the line frequency transformers does not ensure the elimination of DC grid current. The DC component is mainly caused by an offset in the grid current measurement that is usually realized through Hall effect sensors. As known, these sensors are strongly affected by DC offset drift. A current sensor based on fluxgate technology could be adopted to reduce the DC offset drift. The absence of the line frequency transformer also causes two major drawbacks: safety issue in case of isolation faults, and wrong operation because of ground leakage currents. The first issue occurs since the PV generator is not isolated from the grid, as a consequence isolation faults at the PV generator or at DC-wiring can get a safety hazard in case of indirect electric shock. Hence, transformerless inverters are equipped with Residual Current Devices (RCD), which monitor the ground leakage currents during operation. When a dangerous leakage current is recognized, the inverter is immediately disconnected from the AC-grid. A typical configuration for a transformerless power converter, based on H-brige configuration, is reported in Figure 16. Fig. 16. Schematic representation of a transformerless power converter, based on H bridge, for PV systems. Leakage ground currents are shown as dashed lines. The second issue is created by the large parasitic capacitance of the PV panel (10-100nF/kWp) that results in high ground leakage currents in case of variations of common mode voltage at the output of the power converter. Impulsive common mode voltages will result in large ground leakage currents because the power converter is coupled to a resonant circuit with a small damping coefficient. The resonant circuit is composed by the earth capacitance, the converter, the AC filter and the grid. A blocking filter is not easy to tune, since the earth capacitance changes with environmental conditions. The damping of the resonant circuit is small since it is mandatory to reduce dissipative terms and thus to achieve a high efficiency. Therefore the best remedial strategy for ground leakage current is the reduction of the excitation, i.e. of variation of common mode voltage. To this aim it could be necessary to adopt topologies free from common-mode voltage variations (Lopez et al., 2007), (Gonzales et al., 2007). Converter topologies intrinsically safe from leakage currents are the Half-Bridge (HB) and the Neutral Point Clamped (NPC) ones (Gonzales et al., 2008), (Lopez et al., 2006). However, these configurations require twice the input voltage to obtain the same voltage level that is one of the major drawbacks for a PV system. On the other hand, a full-bridge converter driven by bipolar PWM (Gonzales et al 2007) exploits the full voltage rails but reduces common mode voltage with lower efficiency than HB of NPC configurations. From a power quality point of view all these solutions present current ripple at switching frequency. For power converters with transformers the full bridge converter with unipolar PWM is widespread, since it represents a nice trade-off between efficiency, complexity and price. The biggest advantages of unipolar PWM are that the current ripple appears at twice the switching frequency and that the voltage across the AC output filter is unipolar, thus Single-Phase Grid Connected Converters for Photovoltaic Plants 101Renewable Energy102 reducing core losses. In summary, with unipolar PWM the maximum current ripple is four time smaller than the case of bipolar PWM. However, in terms of ground leakage currents the unipolar PWM generates common mode voltage at switching frequency with a peak to peak value equal to the DC voltage Bus. As a consequence, a big common-mode filter is required in photovoltaic transformerless applications relying on full bridge configurations with unipolar PWM. Here an original architecture is proposed where a full bridge converter with a unipolar PWM is used with two additional blocks that ensure no common-mode voltage variations also in case of actual (asymmetrical) commutations. The instantaneous common-mode voltage in the full-bridge converter of the Figure 16 can be computed from the two mid-points of both legs by: 02A BOcmv vv (5) The most widespread single-phase topology is the full bridge one. This topology allows both bipolar and unipolar PWM. In the following sections the behavior of full bridge topologies with bipolar and unipolar PWM will be presented showing performances in terms of ground leakage currents and the original compensation circuitry. 6.1 Full bridge topologies with bipolar PWM for PV transformerless converters Ideally, the full bridge converter with bipolar PWM does not generate a variation of common-mode output voltage. Hence, it is adopted in some commercial transformerless converters. With reference to the schematic of Figure 16 two configurations are sequentially operated in the switching cycle: 1) T1 and T4 ON (T2, T3 OFF): 0 0, 0, 2DCA DC B cmVv V v v   2) T2 and T3 ON (T1,T4 OFF): 0 00, , 2DCA B DC cmVv v V v   In case of ideally synchronous commutation of the switches the common-mode voltage is kept constant at half the DC bus level, and thus no ground leakage currents would appear. However, in actual converters, a small common-mode high-frequency filter is necessary to avoid ground leakage currents due to switching mismatch and asymmetries. Recently modified version of the full bridge configuration with bipolar PWM were proposed in order to combine higher efficiency with reduced ground leakage currents in real operations. Figure 17 presents a modified full-bridge topology with two additional blocks used alternatively: inserting only the DC decoupling block a former topology, known as H5, is obtained while, inserting the AC decoupling block a latter one, known as HERIC, is obtained (Kerekes et al., 2007), (Gonzalez et al., 2006). The use of DC or AC decoupling allows, during conduction of freewheeling diodes, the disconnection of the grid from photovoltaic panel. If no conduction of freewheeling diodes occurs the power switch of DC Decoupling in H5 topology is on while the power switches of the AC decoupling in HERIC topology are off. The H bridge is driven according to 1) and 2) and common mode voltage will be equal to VDC/2. On the other side during conduction of freewheeling diodes H5 topology modifies the PWM involving only the high side of the bridge. DC decoupling is OFF hence, in case of symmetrical commutations 0Av = 0Bv = 2DCV . In this way common mode voltage does not vary avoiding ground leakage currents increase. The HERIC topology obtains the same result turning on one of the two added power switches of the AC decoupling block in function of the sign of the injected grid current and driving off the transistors of the H bridge. So doing the freewheeling will involve only the AC Decoupling. In both topologies the output voltage is zero during freewheeling obtaining a three level output voltage as in case of unipolar modulation. However the output current ripple is still at switching frequency. Again asymmetrical PWM commutations can determine little leakage ground currents. Fig. 17. Full Bridge topology with DC and AC decoupling additional blocks. 6.2 Full bridge topology with unipolar PWM for PV transformerless converters The full bridge topology with unipolar PWM ensures a reduction of the current ripple, specifically the maximum current ripple is four times smaller then the bipolar PWM maximum current ripple. Referring to Figure 16 the common-mode voltage resulting during a switching cycle in case of unipolar PWM is computed in the following. Reference is made to positive output voltage and current (first quadrant operations); output voltage is indicated as 0 0AB A Bv v v  . Four configurations are sequentially operated in the switching cycle: 1) T1, T4 On (T2, T3 OFF): AB DCv V , 2DCcmVv  . 2) T2, T4, D2 ON: 0ABv  . Low side current freewheeling through D2, T4, 0cmv  . 3) T1,T4 On (T2, T3 OFF): AB DCv V , 2DCcmVv  . 4) T1, T3, D3 ON: 0ABv  . High side current freewheeling through T1, D3, cm DCv V . Single-Phase Grid Connected Converters for Photovoltaic Plants 103Renewable Energy102 reducing core losses. In summary, with unipolar PWM the maximum current ripple is four time smaller than the case of bipolar PWM. However, in terms of ground leakage currents the unipolar PWM generates common mode voltage at switching frequency with a peak to peak value equal to the DC voltage Bus. As a consequence, a big common-mode filter is required in photovoltaic transformerless applications relying on full bridge configurations with unipolar PWM. Here an original architecture is proposed where a full bridge converter with a unipolar PWM is used with two additional blocks that ensure no common-mode voltage variations also in case of actual (asymmetrical) commutations. The instantaneous common-mode voltage in the full-bridge converter of the Figure 16 can be computed from the two mid-points of both legs by: 02A BOcmv vv (5) The most widespread single-phase topology is the full bridge one. This topology allows both bipolar and unipolar PWM. In the following sections the behavior of full bridge topologies with bipolar and unipolar PWM will be presented showing performances in terms of ground leakage currents and the original compensation circuitry. 6.1 Full bridge topologies with bipolar PWM for PV transformerless converters Ideally, the full bridge converter with bipolar PWM does not generate a variation of common-mode output voltage. Hence, it is adopted in some commercial transformerless converters. With reference to the schematic of Figure 16 two configurations are sequentially operated in the switching cycle: 1) T1 and T4 ON (T2, T3 OFF): 0 0, 0, 2DCA DC B cmVv V v v   2) T2 and T3 ON (T1,T4 OFF): 0 00, , 2DCA B DC cmVv v V v   In case of ideally synchronous commutation of the switches the common-mode voltage is kept constant at half the DC bus level, and thus no ground leakage currents would appear. However, in actual converters, a small common-mode high-frequency filter is necessary to avoid ground leakage currents due to switching mismatch and asymmetries. Recently modified version of the full bridge configuration with bipolar PWM were proposed in order to combine higher efficiency with reduced ground leakage currents in real operations. Figure 17 presents a modified full-bridge topology with two additional blocks used alternatively: inserting only the DC decoupling block a former topology, known as H5, is obtained while, inserting the AC decoupling block a latter one, known as HERIC, is obtained (Kerekes et al., 2007), (Gonzalez et al., 2006). The use of DC or AC decoupling allows, during conduction of freewheeling diodes, the disconnection of the grid from photovoltaic panel. If no conduction of freewheeling diodes occurs the power switch of DC Decoupling in H5 topology is on while the power switches of the AC decoupling in HERIC topology are off. The H bridge is driven according to 1) and 2) and common mode voltage will be equal to VDC/2. On the other side during conduction of freewheeling diodes H5 topology modifies the PWM involving only the high side of the bridge. DC decoupling is OFF hence, in case of symmetrical commutations 0Av = 0Bv = 2DCV . In this way common mode voltage does not vary avoiding ground leakage currents increase. The HERIC topology obtains the same result turning on one of the two added power switches of the AC decoupling block in function of the sign of the injected grid current and driving off the transistors of the H bridge. So doing the freewheeling will involve only the AC Decoupling. In both topologies the output voltage is zero during freewheeling obtaining a three level output voltage as in case of unipolar modulation. However the output current ripple is still at switching frequency. Again asymmetrical PWM commutations can determine little leakage ground currents. Fig. 17. Full Bridge topology with DC and AC decoupling additional blocks. 6.2 Full bridge topology with unipolar PWM for PV transformerless converters The full bridge topology with unipolar PWM ensures a reduction of the current ripple, specifically the maximum current ripple is four times smaller then the bipolar PWM maximum current ripple. Referring to Figure 16 the common-mode voltage resulting during a switching cycle in case of unipolar PWM is computed in the following. Reference is made to positive output voltage and current (first quadrant operations); output voltage is indicated as 0 0AB A Bv v v  . Four configurations are sequentially operated in the switching cycle: 1) T1, T4 On (T2, T3 OFF): AB DCv V , 2DCcmVv  . 2) T2, T4, D2 ON: 0ABv  . Low side current freewheeling through D2, T4, 0cmv  . 3) T1,T4 On (T2, T3 OFF): AB DCv V , 2DCcmVv  . 4) T1, T3, D3 ON: 0ABv  . High side current freewheeling through T1, D3, cm DCv V . Single-Phase Grid Connected Converters for Photovoltaic Plants 103Renewable Energy104 The common mode voltage vcm varies from 0 to VDC and thus a large ground leakage current will flow. Figure 18 shows the original configuration that is proposed to reduce ground leakage currents. Two additional blocks are added: the former one, named DC decoupling, fixes the common mode voltage to VDC/2, while the latter, named No-Ideality compensation, helps to keep a constant common mode voltage in case of non ideal (asymmetrical) commutations. The DC decoupling disconnects the H-bridge from DC source during freewheeling. In particular during high side freewheeling T5 is switched off while during low side freewheeling T6 is switched off. In case of ideal commutations this block is sufficient to guarantee no variations of the common mode voltage cmv . In actual converter, to fix the cmv at 2DCV during asymmetrical commutations, two additional low power switches should be added. In order to avoid the complexity introduced by two controlled switches an attempt was made replacing them with two diodes. The idea relies on the following explanation. Once the decoupling switch T5 opens, the voltage of the high side of the H bridge, is floating. Then this voltage, due to parasitic capacitances, would decrease, however the inserted diode will clamp it to VDC/2. Simulation results confirm this behavior. Since the two diodes impose a maximum voltage equal to VDC/2 across T5 and T6, latest generation Mosfets can be chosen as controlled switches, in order to reduce the power losses. In figure 18 x and y represent the PWM signals used for driving the legs of the full bridge. A simple combination of PWM signals can be used for driving the switches of DC decoupling block. With reference to the schematic of Figure 18 four configuration are sequentially operated in the switching cycle: 1) T1, T4, T5,T6 ON (T2 and T3 OFF): AB DCv V , 2DCcmVv  . 2) T2, T4, T5, D2 ON (T1, T3, T6 OFF): 0ABv  . Low side current freewheeling through T4, D4, 0 0 2DCcm A BVv v v   . 3) T1, T4, T5, T6 ON (T2 and T3 OFF): AB DCv V , 2DCcmVv  . 4) T1, T3, T6, D3 ON (T2, T4, T5 OFF): 0ABv  . High side current freewheeling through T1,D3, 0 0 2DCcm A BVv v v   . Fig. 18. Schematic of the original configuration for full bridge topology with unipolar PWM for transformerless converters. In an actual converter the PWM modulation strategy of the DC decoupling block has to be modified because of the presence of dead times. A simple adaptation of the driving circuit can be made that ensures correct operations in actual conditions. 6.3 Simulation Results The proposed topology (Figure 18) was validated thanks to the PLECS toolbox, which allows fast simulation of power electronic circuits under MATLAB/Simulink. The simulation parameters are: 450DCV V , 230grid rmsV V , 2groundR   , 14PVgroundC nF , 2fL mH , C f  2.2F , 0.5fR   , Lgrid  40H , the switching frequency is 10swf kHz . A simple current controller is used to inject a sinusoidal current of 13Arms into the grid with unitary power factor equivalent to 3kW of injected active power. Extensive simulations were performed with the attempt to assess the performances of the proposed converter topology with special reference to common mode voltage and ground leakage currents. Results confirm that the proposed topology allows to keep the common voltage constant at VDC/2. In this way, the resulting ground leakage currents are negligible. Figures 19, 20, 21 show respectively the common mode voltage cmv , the ground voltage groundv and the ground leakage currents cmi without DC decoupling and No-Ideality compensation additional blocks. These simulations were made to quantify the common mode voltage and the ground leakage currents arising in a transformerless full bridge topology with unipolar PWM. Figures 22, 23 show the simulation results obtained when the two additional blocks are added. Figure 22 shows that the common mode voltage cmv is fixed to 2DCV . Therefore the ground voltage groundv does not contain high frequency components and only the fundamental grid frequency is present (Figure 22). The resulting leakage ground currents will be very small, as shown in Figure 23. Single-Phase Grid Connected Converters for Photovoltaic Plants 105Renewable Energy104 The common mode voltage vcm varies from 0 to VDC and thus a large ground leakage current will flow. Figure 18 shows the original configuration that is proposed to reduce ground leakage currents. Two additional blocks are added: the former one, named DC decoupling, fixes the common mode voltage to VDC/2, while the latter, named No-Ideality compensation, helps to keep a constant common mode voltage in case of non ideal (asymmetrical) commutations. The DC decoupling disconnects the H-bridge from DC source during freewheeling. In particular during high side freewheeling T5 is switched off while during low side freewheeling T6 is switched off. In case of ideal commutations this block is sufficient to guarantee no variations of the common mode voltage cmv . In actual converter, to fix the cmv at 2DCV during asymmetrical commutations, two additional low power switches should be added. In order to avoid the complexity introduced by two controlled switches an attempt was made replacing them with two diodes. The idea relies on the following explanation. Once the decoupling switch T5 opens, the voltage of the high side of the H bridge, is floating. Then this voltage, due to parasitic capacitances, would decrease, however the inserted diode will clamp it to VDC/2. Simulation results confirm this behavior. Since the two diodes impose a maximum voltage equal to VDC/2 across T5 and T6, latest generation Mosfets can be chosen as controlled switches, in order to reduce the power losses. In figure 18 x and y represent the PWM signals used for driving the legs of the full bridge. A simple combination of PWM signals can be used for driving the switches of DC decoupling block. With reference to the schematic of Figure 18 four configuration are sequentially operated in the switching cycle: 1) T1, T4, T5,T6 ON (T2 and T3 OFF): AB DCv V , 2DCcmVv  . 2) T2, T4, T5, D2 ON (T1, T3, T6 OFF): 0ABv  . Low side current freewheeling through T4, D4, 0 0 2DCcm A BVv v v   . 3) T1, T4, T5, T6 ON (T2 and T3 OFF): AB DCv V , 2DCcmVv  . 4) T1, T3, T6, D3 ON (T2, T4, T5 OFF): 0ABv  . High side current freewheeling through T1,D3, 0 0 2DCcm A BVv v v   . Fig. 18. Schematic of the original configuration for full bridge topology with unipolar PWM for transformerless converters. In an actual converter the PWM modulation strategy of the DC decoupling block has to be modified because of the presence of dead times. A simple adaptation of the driving circuit can be made that ensures correct operations in actual conditions. 6.3 Simulation Results The proposed topology (Figure 18) was validated thanks to the PLECS toolbox, which allows fast simulation of power electronic circuits under MATLAB/Simulink. The simulation parameters are: 450DCV V , 230grid rmsV V , 2groundR   , 14PVgroundC nF , 2fL mH , C f  2.2F , 0.5fR   , Lgrid  40H , the switching frequency is 10swf kHz . A simple current controller is used to inject a sinusoidal current of 13Arms into the grid with unitary power factor equivalent to 3kW of injected active power. Extensive simulations were performed with the attempt to assess the performances of the proposed converter topology with special reference to common mode voltage and ground leakage currents. Results confirm that the proposed topology allows to keep the common voltage constant at VDC/2. In this way, the resulting ground leakage currents are negligible. Figures 19, 20, 21 show respectively the common mode voltage cmv , the ground voltage groundv and the ground leakage currents cmi without DC decoupling and No-Ideality compensation additional blocks. These simulations were made to quantify the common mode voltage and the ground leakage currents arising in a transformerless full bridge topology with unipolar PWM. Figures 22, 23 show the simulation results obtained when the two additional blocks are added. Figure 22 shows that the common mode voltage cmv is fixed to 2DCV . Therefore the ground voltage groundv does not contain high frequency components and only the fundamental grid frequency is present (Figure 22). The resulting leakage ground currents will be very small, as shown in Figure 23. Single-Phase Grid Connected Converters for Photovoltaic Plants 105Renewable Energy106 Fig. 19. Simulation results. Common mode voltage in a transformerless unipolar PWM full bridge converter. Fig. 20. Simulation results. Ground voltage in a transformerless unipolar PWM full bridge converter. Fig. 21. Simulation results. Ground leakage current in a transformerless unipolar PWM full bridge converter. Fig. 22. Simulation results. Common mode voltage (red trace) and ground voltage (blue trace) with the proposed topology. Fig. 23. Simulation results. Ground leakage currents [mA] with the proposed topology. 7. Conclusion This chapter deals with single-phase grid connected converters used in photovoltaic plants. Some proposed improvements to the widespread full bridge topology are investigated in order to improve energy conversion. PV power converters, based on full bridge topology, are distinguished by the presence or not of a line frequency transformer. In both cases the control of the injected grid current is one of the most important issue. The d-q current controller was deeply analyzed and some experimental results are reported to show its superior power quality in comparison with traditional controllers. In order to increase the performances, notably at light loads conditions, the full bridge topology was modified moving the feedback from the transformer input to the transformer output. In this way, the DC component at the transformer input is not directly controlled, thus an original core compensation strategy is proposed to avoid saturation. Experimental results are reported showing the effectiveness of the proposed compensation. In low power applications the trend is to remove the transformer (transformerless configurations) in order to reduce losses, costs and size. In transformerless configurations the major issues are the accurate control of DC component of injected grid current, the safety issues in case of isolation faults, and the reduction of ground leakage currents. Here the most common configurations are reviewed and an original solution is presented that achieves a nice trade-off between efficiency, power quality and reduction of ground leakage currents. 8. References Chung, S.K. (2000). Phase-Locked Loop for Grid-Connected Three-Phase power conversion systems, IEE Proc. – Electr. Power Appl., Vol. 147, No. 3, pp. 213-219. Arruda, L. N.; Cardoso Filho, B. J.; Silva, S. M. & Diniz, S.A.C. (2001). Wide bandwidth single and three-phase PLL structures for grid-tied PV systems, in Proc. of EPE’01, pp. 1660-1663. Silva, S.M.; Filho, B.C.; Campana, R.P. & Boaventura, W.C (2004). Performance evaluation of PLL algorithms for single- phase grid-connected systems, in Proc. of IAS’04, vol. 4, pp. 2259 - 2263. C. J. Kaufman, Rocky Mountain Research Lab. Yuan, X.; Merk, W.; Stemmler, H. & Allmeling, J. (2002). Stationary-frame generalized integrators for current control of active power filters with zero steady-state error for current harmonics of concern under unbalanced and distorted operating conditions. IEEE Transactions on Industry Applications, vol. 38, no. 2, pp. 523–532, Mar./Apr. 2002. Teodorescu, R.; Blaabjerg, F.; Borup, U. & Liserre M. (2004). A new control structure for grid-connected LCL PV inverters with zero steady-state error and selective harmonic compensation, in Proc. of Applied Power Electronics Conference and Exposition, 2004. APEC ’04. Nineteenth Annual IEEE, vol. 1, pp. 580–586. Ryan, M. J. & Lorenz R. D. (1997). A synchronous-frame controller for a single-phase sine waveinverter. in Proc. Of Applied Power Electronics Conference and Exposition. APEC ’97 Conference Proceedings 1997., Twelfth Annual, vol. 2, Atlanta, GA, USA, Feb. 1997, pp. 813–819. Miranda, U. A.; Rolim, L. G. B. & Aredes M. (2005). A DQ synchronous reference frame current control for single-phase converters, in Proc. Of Power Electronics Specialists Conference. PESC ’05. IEEE 36th, 2005, pp. 1377–1381. Cacciato, M.; Consoli, A.; Aiello, N.; Attanasio, R.; Gennaro, F. & Macina G. (2008). A digitally controlled double stage soft-switching converter for grid-connected photovoltaic applications, in Proc. Of Applied Power Electronics Conference and Exposition. APEC 2008. Twenty-Third Annual IEEE, Feb. 2008, pp. 141–147. Bellini, A.; Franceschini, G.; Lorenzani, E. & Tassoni, C. (2006). Synchronous reference frame grid current control for single-phase photovoltaic converters, in Proc. of IAS 2008, pp 1 – 7, Edmonton (Canada), October 5-9 2008 Ciobotaru, M.; Teodorescu, R. & Blaabjerg, F. (2005). Control of single-stage single-phase PV inverter, in Proc. of Power Electronics and Applications, 2005 European Conference on, Sept. 2005. Calais, M. & Agelidis, V. G. (1998). Multilevel converters for single-phase grid connected photovoltaic systems—An overview, in Proc. IEEE Int. Symp. Ind. Electron., vol. 1, pp. 224–229. Calais, M.; Myrzik, J. M. A. & Agelidis, V. G. (2001). Inverters for single phase grid connected photovoltaic systems—Overview and prospects, in Proc. 17th Eur. Photovoltaic Solar Energy Conf., Munich, Germany, Oct. 22–26, pp. 437–440. Single-Phase Grid Connected Converters for Photovoltaic Plants 107Renewable Energy106 Fig. 19. Simulation results. Common mode voltage in a transformerless unipolar PWM full bridge converter. Fig. 20. Simulation results. Ground voltage in a transformerless unipolar PWM full bridge converter. Fig. 21. Simulation results. Ground leakage current in a transformerless unipolar PWM full bridge converter. Fig. 22. Simulation results. Common mode voltage (red trace) and ground voltage (blue trace) with the proposed topology. Fig. 23. Simulation results. Ground leakage currents [mA] with the proposed topology. 7. Conclusion This chapter deals with single-phase grid connected converters used in photovoltaic plants. Some proposed improvements to the widespread full bridge topology are investigated in order to improve energy conversion. PV power converters, based on full bridge topology, are distinguished by the presence or not of a line frequency transformer. In both cases the control of the injected grid current is one of the most important issue. The d-q current controller was deeply analyzed and some experimental results are reported to show its superior power quality in comparison with traditional controllers. In order to increase the performances, notably at light loads conditions, the full bridge topology was modified moving the feedback from the transformer input to the transformer output. In this way, the DC component at the transformer input is not directly controlled, thus an original core compensation strategy is proposed to avoid saturation. Experimental results are reported showing the effectiveness of the proposed compensation. In low power applications the trend is to remove the transformer (transformerless configurations) in order to reduce losses, costs and size. In transformerless configurations the major issues are the accurate control of DC component of injected grid current, the safety issues in case of isolation faults, and the reduction of ground leakage currents. Here the most common configurations are reviewed and an original solution is presented that achieves a nice trade-off between efficiency, power quality and reduction of ground leakage currents. 8. References Chung, S.K. (2000). Phase-Locked Loop for Grid-Connected Three-Phase power conversion systems, IEE Proc. – Electr. Power Appl., Vol. 147, No. 3, pp. 213-219. Arruda, L. N.; Cardoso Filho, B. J.; Silva, S. M. & Diniz, S.A.C. (2001). Wide bandwidth single and three-phase PLL structures for grid-tied PV systems, in Proc. of EPE’01, pp. 1660-1663. Silva, S.M.; Filho, B.C.; Campana, R.P. & Boaventura, W.C (2004). Performance evaluation of PLL algorithms for single- phase grid-connected systems, in Proc. of IAS’04, vol. 4, pp. 2259 - 2263. C. J. Kaufman, Rocky Mountain Research Lab. Yuan, X.; Merk, W.; Stemmler, H. & Allmeling, J. (2002). Stationary-frame generalized integrators for current control of active power filters with zero steady-state error for current harmonics of concern under unbalanced and distorted operating conditions. IEEE Transactions on Industry Applications, vol. 38, no. 2, pp. 523–532, Mar./Apr. 2002. Teodorescu, R.; Blaabjerg, F.; Borup, U. & Liserre M. (2004). A new control structure for grid-connected LCL PV inverters with zero steady-state error and selective harmonic compensation, in Proc. of Applied Power Electronics Conference and Exposition, 2004. APEC ’04. Nineteenth Annual IEEE, vol. 1, pp. 580–586. Ryan, M. J. & Lorenz R. D. (1997). A synchronous-frame controller for a single-phase sine waveinverter. in Proc. Of Applied Power Electronics Conference and Exposition. APEC ’97 Conference Proceedings 1997., Twelfth Annual, vol. 2, Atlanta, GA, USA, Feb. 1997, pp. 813–819. Miranda, U. A.; Rolim, L. G. B. & Aredes M. (2005). A DQ synchronous reference frame current control for single-phase converters, in Proc. Of Power Electronics Specialists Conference. PESC ’05. IEEE 36th, 2005, pp. 1377–1381. Cacciato, M.; Consoli, A.; Aiello, N.; Attanasio, R.; Gennaro, F. & Macina G. (2008). A digitally controlled double stage soft-switching converter for grid-connected photovoltaic applications, in Proc. Of Applied Power Electronics Conference and Exposition. APEC 2008. Twenty-Third Annual IEEE, Feb. 2008, pp. 141–147. Bellini, A.; Franceschini, G.; Lorenzani, E. & Tassoni, C. (2006). Synchronous reference frame grid current control for single-phase photovoltaic converters, in Proc. of IAS 2008, pp 1 – 7, Edmonton (Canada), October 5-9 2008 Ciobotaru, M.; Teodorescu, R. & Blaabjerg, F. (2005). Control of single-stage single-phase PV inverter, in Proc. of Power Electronics and Applications, 2005 European Conference on, Sept. 2005. Calais, M. & Agelidis, V. G. (1998). Multilevel converters for single-phase grid connected photovoltaic systems—An overview, in Proc. IEEE Int. Symp. Ind. Electron., vol. 1, pp. 224–229. Calais, M.; Myrzik, J. M. A. & Agelidis, V. G. (2001). Inverters for single phase grid connected photovoltaic systems—Overview and prospects, in Proc. 17th Eur. Photovoltaic Solar Energy Conf., Munich, Germany, Oct. 22–26, pp. 437–440. Single-Phase Grid Connected Converters for Photovoltaic Plants 107Renewable Energy108 Kuo, Y. C.; Liang, T. J. & Chen, J. F. (2001). Novel maximum-power-pointtracking controller for photovoltaic energy conversion system, IEEE Trans. Ind. Electron., vol. 48, no. 3, pp. 594–601, Jun. 2001. Femia, N.; Petrone, G.; Spagnuolo, G. & Vitelli, M. (2005). Optimization of perturb and observe maximum power point tracking method, Power Electronics, IEEE Transactions on , vol.20, no.4, pp. 963-973, July 2005 Fangrui Liu; Shanxu Duan; Fei Liu; Bangyin Liu & Yong Kang,(2008). A Variable Step Size INC MPPT Method for PV Systems, IEEE Transactions on Industrial Electronics, vol. 55, no. 7, pp. 2622-2628, July 2008 Esram, T.; Kimball, J.W.; Krein, P.T.; Chapman, P.L. & Midya, P. (2006). Dynamic Maximum Power Point Tracking of Photovoltaic Arrays Using Ripple Correlation Control, IEEE Transactions on Power Electronics, vol. 21, no.5, pp. 1282-1291, Sept. 2006 Lopez, Oscar; Teodorescu, Remus; Freijedo, Francisco & Doval-Gandoy, Jesus (2007). Leakage current evaluation of a singlephase transformerless PV inverter connected to the grid, Applied Power Electronics Conference, APEC 2007 - Twenty Second Annual IEEE ,pp. 907 – 912, Feb. 25 2007-March 1 2007 Gonzalez, R.; Lopez, J.; Sanchis, P. & Marroyo, L. (2007). Transformerless Inverter for Single-Phase Photovoltaic Systems, IEEE Transactions on Power Electronics Volume 22, Issue 2, March 2007, pp. 693 – 697. Gonzalez, R.; Gubia, E.; Lopez, J. & Marroyo, L. (2008). Transformerless Single-Phase Multilevel-Based Photovoltaic Inverter, IEEE Transactions on Industrial Electronics Volume 55, Issue 7, July 2008, pp. 2694 – 2702. Lopez, Oscar; Teodorescu, Remus & Doval-Gandoy, Jesus.(2006). Multilevel transformerless topologies for single-phase grid-connected converters, IEEE IECON 2006 - 32nd Annual Conference on Nov. 2006, pp. 5191 – 5196. Kerekes, T.; Teodorescu & R.; Borup, U. (2007). Transformerless Photovoltaic Inverters Connected to the Grid, Applied Power Electronics Conference, APEC 2007 - Twenty Second Annual IEEE Feb. 25 2007-March 1 2007, pp.1733 – 1737. Gonzalez, R.; Lopez, J.; Sanchis, P.; Gubia, E.; Ursua, A. & Marroyo, L. (2006). High-Efficiency Transformerless Single-phase Photovoltaic Inverter, Power Electronics and Motion Control Conference, 2006. EPE-PEMC 2006, 12th International Aug. 2006, pp. 1895 – 1900. Grid Integration of Renewable Energy Systems 109Grid Integration of Renewable Energy SystemsAthula Rajapakse, Dharshana Muthumuni and Nuwan PereraX Grid Integration of Renewable Energy Systems Athula Rajapakse1, Dharshana Muthumuni2 and Nuwan Perera1 University of Manitoba1 Manitoba HVDC Research Centre2 Manitoba, Canada 1. Introduction Integration of renewable energy into the utility grid can be at either the transmission level or the distribution level, depending on the scale of generation. Large renewable energy generation such as wind farms are directly interconnected to the transmission system. Small scale distributed generation is generally interconnected to the medium or low voltage distribution systems. Both types of interconnections present different challenges that must be carefully analysed before systems are designed. This chapter outlines and discusses the most common issues encountered during grid integration of different types of renewable energy systems. The application of detailed simulation methods to identify problems and to find suitable solutions is also highlighted. Few case studies and simulation examples are presented to reinforce the concepts presented in the chapter. 2. Integration of small scale generation into distribution grids Electricity generation using renewable energy resources is often taking place in small scale due to disperse nature of the resources. Good examples are small hydro, solar photovoltaics, biogas, biomass and small wind turbine based electricity generation systems. The size of these generators typically varies from a few hundreds of kilowatts to several megawatts. These small scale electricity generators are generally connected to the grid at the primary or secondary distribution level and are considered distributed generation (DG) or distributed resources (DR). Distributed resources include both renewable and non-renewable small scale generation as well as energy storage. 2.1 Different types of grid interfaces Often, the small scale renewable generators are not directly connected to the grid. The generation technology or the operational characteristics requires the use of some interface between the generator and the utility distribution grid. For example, solar photovoltaics (PV) panels generate dc electricity and therefore, a power electronics based dc-to-ac converter is required between the grid and the generator. Some technologies such as induction generator based small hydro or wind can be directly connected to the ac grid. However, concerns such as starting transients, energy conversion efficiency and power 7Renewable Energy108 Kuo, Y. C.; Liang, T. J. & Chen, J. F. (2001). Novel maximum-power-pointtracking controller for photovoltaic energy conversion system, IEEE Trans. Ind. Electron., vol. 48, no. 3, pp. 594–601, Jun. 2001. Femia, N.; Petrone, G.; Spagnuolo, G. & Vitelli, M. (2005). Optimization of perturb and observe maximum power point tracking method, Power Electronics, IEEE Transactions on , vol.20, no.4, pp. 963-973, July 2005 Fangrui Liu; Shanxu Duan; Fei Liu; Bangyin Liu & Yong Kang,(2008). A Variable Step Size INC MPPT Method for PV Systems, IEEE Transactions on Industrial Electronics, vol. 55, no. 7, pp. 2622-2628, July 2008 Esram, T.; Kimball, J.W.; Krein, P.T.; Chapman, P.L. & Midya, P. (2006). Dynamic Maximum Power Point Tracking of Photovoltaic Arrays Using Ripple Correlation Control, IEEE Transactions on Power Electronics, vol. 21, no.5, pp. 1282-1291, Sept. 2006 Lopez, Oscar; Teodorescu, Remus; Freijedo, Francisco & Doval-Gandoy, Jesus (2007). Leakage current evaluation of a singlephase transformerless PV inverter connected to the grid, Applied Power Electronics Conference, APEC 2007 - Twenty Second Annual IEEE ,pp. 907 – 912, Feb. 25 2007-March 1 2007 Gonzalez, R.; Lopez, J.; Sanchis, P. & Marroyo, L. (2007). Transformerless Inverter for Single-Phase Photovoltaic Systems, IEEE Transactions on Power Electronics Volume 22, Issue 2, March 2007, pp. 693 – 697. Gonzalez, R.; Gubia, E.; Lopez, J. & Marroyo, L. (2008). Transformerless Single-Phase Multilevel-Based Photovoltaic Inverter, IEEE Transactions on Industrial Electronics Volume 55, Issue 7, July 2008, pp. 2694 – 2702. Lopez, Oscar; Teodorescu, Remus & Doval-Gandoy, Jesus.(2006). Multilevel transformerless topologies for single-phase grid-connected converters, IEEE IECON 2006 - 32nd Annual Conference on Nov. 2006, pp. 5191 – 5196. Kerekes, T.; Teodorescu & R.; Borup, U. (2007). Transformerless Photovoltaic Inverters Connected to the Grid, Applied Power Electronics Conference, APEC 2007 - Twenty Second Annual IEEE Feb. 25 2007-March 1 2007, pp.1733 – 1737. Gonzalez, R.; Lopez, J.; Sanchis, P.; Gubia, E.; Ursua, A. & Marroyo, L. (2006). High-Efficiency Transformerless Single-phase Photovoltaic Inverter, Power Electronics and Motion Control Conference, 2006. EPE-PEMC 2006, 12th International Aug. 2006, pp. 1895 – 1900. Grid Integration of Renewable Energy Systems 109Grid Integration of Renewable Energy SystemsAthula Rajapakse, Dharshana Muthumuni and Nuwan PereraX Grid Integration of Renewable Energy Systems Athula Rajapakse1, Dharshana Muthumuni2 and Nuwan Perera1 University of Manitoba1 Manitoba HVDC Research Centre2 Manitoba, Canada 1. Introduction Integration of renewable energy into the utility grid can be at either the transmission level or the distribution level, depending on the scale of generation. Large renewable energy generation such as wind farms are directly interconnected to the transmission system. Small scale distributed generation is generally interconnected to the medium or low voltage distribution systems. Both types of interconnections present different challenges that must be carefully analysed before systems are designed. This chapter outlines and discusses the most common issues encountered during grid integration of different types of renewable energy systems. The application of detailed simulation methods to identify problems and to find suitable solutions is also highlighted. Few case studies and simulation examples are presented to reinforce the concepts presented in the chapter. 2. Integration of small scale generation into distribution grids Electricity generation using renewable energy resources is often taking place in small scale due to disperse nature of the resources. Good examples are small hydro, solar photovoltaics, biogas, biomass and small wind turbine based electricity generation systems. The size of these generators typically varies from a few hundreds of kilowatts to several megawatts. These small scale electricity generators are generally connected to the grid at the primary or secondary distribution level and are considered distributed generation (DG) or distributed resources (DR). Distributed resources include both renewable and non-renewable small scale generation as well as energy storage. 2.1 Different types of grid interfaces Often, the small scale renewable generators are not directly connected to the grid. The generation technology or the operational characteristics requires the use of some interface between the generator and the utility distribution grid. For example, solar photovoltaics (PV) panels generate dc electricity and therefore, a power electronics based dc-to-ac converter is required between the grid and the generator. Some technologies such as induction generator based small hydro or wind can be directly connected to the ac grid. However, concerns such as starting transients, energy conversion efficiency and power 7Renewable Energy110 quality issues make connecting them through a power electronics interface a better choice. Table-1 summarizes some of the common types of generation and their preferred interfacing technologies. Type Interfacing technology Fuel cells Power electronic converter Wind Induction generator/power electronic converter d Photovoltaics Power electronic converter Small hydro Synchronous or induction generator, power electronic converter Table 1. Interfacing technologies 2.2 Issues related to grid integration of small scale generation Most electric distribution systems are designed, operated and protected on the premise of there being a single voltage source on each distribution feeder. Interconnection of small scale renewable generation to the distribution grid violates this fundamental assumption. Therefore, certain special requirements need to be satisfied when interconnecting distributed generation to the grid in order to ensure safe and reliable operation. Fault clearing, reclosing and inadvertent islanded operation are major protection related concerns [IEEE, 2004]. In many cases, the grid-DG interface is based on power electronics inverters or asynchronous generators. The utilities are concerned about their impact on power quality. The impacts include harmonics, voltage dips, overvoltages, and voltage flicker [IEEE, 2003]. Several studies [Aldferfer et al, 2000; Dugan & McDermott, 2002] have identified that the technical requirements imposed by utilities to address these concerns as a major technical barrier for grid integration of DG. The standards such as IEEE P1547 [IEEE, 2003] provide guidelines for developing the general technical requirements for interconnection of small scale generation. 2.2.1 Protection issues Protection is one of the most problematic issues related to the interconnection of distributed generation. Radial systems are the most common form of distribution network configuration. They are usually protected using time graded overcurrent protection schemes. Interconnection of DGs may alter the coordination of the existing protection schemes. If not carefully addressed, this can lead to malfunction of protection equipment. The main protection issues related to the interconnection of DG are briefly discussed below. (a) Change of Short Circuit Levels Short circuit level is one of the main parameters used in the selection of circuit breakers, reclosers, fuses and current transformers (CTs), and the coordination between overcurrent relays. Short circuit level is characterized by the equivalent system impedance at the fault point and indicates the expected level of the fault current. Time variation of the fault current is influenced by the characteristics of the rotating machinery in the vicinity. Most distribution systems are initially designed as passive networks without DG. However, with the connection of DG, equivalent network impedance can decrease, resulting in an increase in the fault level. Thus, in case of a fault, there can be unexpectedly high fault currents which may exceed the interrupting capacity of existing circuit breakers. High fault currents can also lead to CT saturation. Also, the changed fault levels can disrupt the coordination between overcurrent relays leading to unsatisfactory operation of protection systems. To illustrate the concerns, consider the distribution system in Figure 1. When the DG is not present, the fault currents seen by the feeder breaker BK and the recloser RC are approximately equal to the fault current If (Is  Ir  If). However, when the DG is connected, Ir = Is+Idg, and Ir > Is. This is a condition normally not observed in passive radial networks. This doesn’t cause a problem if the recloser interrupting capacity is sufficient to handle the increased fault current. However, it is likely that coordination between the recloser and any downstream fuses (for example between RC and FS1) is lost. Because both the recloser and fuses operate faster at higher fault currents, the required margins between the recloser fast curve and the fuse minimum melt curve could be reduced to an extent that the coordination is lost. Fig. 1. Change of fault currents due to DG interconnection (b) Reverse Power Flow In radial distribution systems, the power flows are unidirectional. Protection schemes are initially designed based on these power flows. When a DG is connected, the power flows may be reversed [Lopes, 2002]. This could also alter the coordination of protection relays. (c) Lack of Sustained Fault Current In order for the protection relays to reliably detect and discriminate fault currents from the normal load currents, the faults must cause a significant and sustained increase in the currents measured by the relays. If the fault current contribution from a DG is limited, it becomes difficult for the overcurrent based protection relays to effectively detect faults. Renewable energy based generation often employs induction generators, small synchronous generators or power electronic converters. Induction generators cannot supply sustained fault currents to three-phase faults and make only a limited fault current contribution to asymmetrical faults. Small synchronous generators are usually not able to supply sustained fault currents that are significantly greater than the rated current. Power semiconductor devices cannot withstand significant overcurrent for sustained periods and therefore power electronic converters are designed to internally limit the output current. The lack of sustained fault current compromises the ability of relays to detect faults. [Barker & de Mello, 2000]. (d) Islanding Islanding is the situation where a part of the utility network is disconnected from the main grid and operates as an independent system supplied with one or more generators. Islanding results in abnormal variations of frequency and voltage in the ‘island’. Opening of an auto recloser during a fault may lead to the formation of two independent systems that Grid Ir IS DG Idg If FaultFS3RCBK FS2FS1Grid Integration of Renewable Energy Systems 111Renewable Energy110 quality issues make connecting them through a power electronics interface a better choice. Table-1 summarizes some of the common types of generation and their preferred interfacing technologies. Type Interfacing technology Fuel cells Power electronic converter Wind Induction generator/power electronic converter d Photovoltaics Power electronic converter Small hydro Synchronous or induction generator, power electronic converter Table 1. Interfacing technologies 2.2 Issues related to grid integration of small scale generation Most electric distribution systems are designed, operated and protected on the premise of there being a single voltage source on each distribution feeder. Interconnection of small scale renewable generation to the distribution grid violates this fundamental assumption. Therefore, certain special requirements need to be satisfied when interconnecting distributed generation to the grid in order to ensure safe and reliable operation. Fault clearing, reclosing and inadvertent islanded operation are major protection related concerns [IEEE, 2004]. In many cases, the grid-DG interface is based on power electronics inverters or asynchronous generators. The utilities are concerned about their impact on power quality. The impacts include harmonics, voltage dips, overvoltages, and voltage flicker [IEEE, 2003]. Several studies [Aldferfer et al, 2000; Dugan & McDermott, 2002] have identified that the technical requirements imposed by utilities to address these concerns as a major technical barrier for grid integration of DG. The standards such as IEEE P1547 [IEEE, 2003] provide guidelines for developing the general technical requirements for interconnection of small scale generation. 2.2.1 Protection issues Protection is one of the most problematic issues related to the interconnection of distributed generation. Radial systems are the most common form of distribution network configuration. They are usually protected using time graded overcurrent protection schemes. Interconnection of DGs may alter the coordination of the existing protection schemes. If not carefully addressed, this can lead to malfunction of protection equipment. The main protection issues related to the interconnection of DG are briefly discussed below. (a) Change of Short Circuit Levels Short circuit level is one of the main parameters used in the selection of circuit breakers, reclosers, fuses and current transformers (CTs), and the coordination between overcurrent relays. Short circuit level is characterized by the equivalent system impedance at the fault point and indicates the expected level of the fault current. Time variation of the fault current is influenced by the characteristics of the rotating machinery in the vicinity. Most distribution systems are initially designed as passive networks without DG. However, with the connection of DG, equivalent network impedance can decrease, resulting in an increase in the fault level. Thus, in case of a fault, there can be unexpectedly high fault currents which may exceed the interrupting capacity of existing circuit breakers. High fault currents can also lead to CT saturation. Also, the changed fault levels can disrupt the coordination between overcurrent relays leading to unsatisfactory operation of protection systems. To illustrate the concerns, consider the distribution system in Figure 1. When the DG is not present, the fault currents seen by the feeder breaker BK and the recloser RC are approximately equal to the fault current If (Is  Ir  If). However, when the DG is connected, Ir = Is+Idg, and Ir > Is. This is a condition normally not observed in passive radial networks. This doesn’t cause a problem if the recloser interrupting capacity is sufficient to handle the increased fault current. However, it is likely that coordination between the recloser and any downstream fuses (for example between RC and FS1) is lost. Because both the recloser and fuses operate faster at higher fault currents, the required margins between the recloser fast curve and the fuse minimum melt curve could be reduced to an extent that the coordination is lost. Fig. 1. Change of fault currents due to DG interconnection (b) Reverse Power Flow In radial distribution systems, the power flows are unidirectional. Protection schemes are initially designed based on these power flows. When a DG is connected, the power flows may be reversed [Lopes, 2002]. This could also alter the coordination of protection relays. (c) Lack of Sustained Fault Current In order for the protection relays to reliably detect and discriminate fault currents from the normal load currents, the faults must cause a significant and sustained increase in the currents measured by the relays. If the fault current contribution from a DG is limited, it becomes difficult for the overcurrent based protection relays to effectively detect faults. Renewable energy based generation often employs induction generators, small synchronous generators or power electronic converters. Induction generators cannot supply sustained fault currents to three-phase faults and make only a limited fault current contribution to asymmetrical faults. Small synchronous generators are usually not able to supply sustained fault currents that are significantly greater than the rated current. Power semiconductor devices cannot withstand significant overcurrent for sustained periods and therefore power electronic converters are designed to internally limit the output current. The lack of sustained fault current compromises the ability of relays to detect faults. [Barker & de Mello, 2000]. (d) Islanding Islanding is the situation where a part of the utility network is disconnected from the main grid and operates as an independent system supplied with one or more generators. Islanding results in abnormal variations of frequency and voltage in the ‘island’. Opening of an auto recloser during a fault may lead to the formation of two independent systems that Grid Ir IS DG Idg If FaultFS3RCBK FS2FS1Grid Integration of Renewable Energy Systems 111Renewable Energy112 operate at two different frequencies [Jenkins et al, 2000]. Reclosing of the auto recloser while the two systems are out of phase could bring disastrous results. [Borbely & Kreider, 2000]. Further, islanding operation may create an ungrounded system depending on the transformer connection. An unidentified island would be hazardous for the repair crews. Due to the above reasons, islanding is recognized as an unsafe situation and immediate disconnection of the DGs from the grid is recommended upon formation of an island [IEEE, 2003]. 2.2.2 Voltage control Interconnection of DG results in changes in power flows and the voltage profile of the feeder, and generally results in overvoltages under low load or high (DG) production conditions. In weak networks, the DG capacity is generally determined by the voltage limits. Furthermore, the connection status of a DG is not controlled by the utility. Disconnection of a DG during the high load can cause undervoltages, while re-connection of a DG under low load conditions may cause overvoltages. This may lead to poor power quality situation and may result in operation of under/over voltage relays. Selection of appropriate tap settings for the distribution transformers becomes difficult with the increased penetration of DG. This is specially difficult when the DGs are not equally distributed among the feeders supplied by the same transformer [Uchida et al, 2006]. Such a situation is illustrated in Figure 2 where there are two feeders supplied by the same transformer, but DGs are concentrated on only one of them. Generally, when the DGs are connected, the net current flow through the transformer is reduced because the DGs provide the power to the nearby loads. Consequently, the transformer tap needs to be changed to the light load setting. The resulting decline of the sending voltage can cause a voltage violation at the far end of the feeder without DGs as shown in Figure 2. Leaving the transformer tap at the heavy load setting risks overvoltages on the feeder with DGs. Switched capacitors and static VAr compensators can be used to control the feeder voltages, but these solutions are often too costly. Fig. 2. Possible DG interconnection configurations Grid DG DGVoltage magnitude Heavy load settingDistribution distance VmaxLight load settingVmin Feeder with DG Feeder without DGFeeder with DG Feeder without DGVoltage limit violations Another issue that affects the operation of DGs at distribution level is the unbalanced voltage profile. As shown in Figure 3, both the loads as well as DGs can be either three-phase or single-phase. Interconnection of single phase sources will increase the system unbalance. On the other hand, inherently unbalanced distribution systems can pose problems for the three-phase DGs connected to it: the resulting unbalance currents in the DG can cause overheating and frequent shutdowns. Fig. 3. Possible DG interconnection configurations 2.2.3 Harmonics and flicker The quality of power supplied to the customers is one of the major concerns of service providers. The majority of the DGs are interfaced using power electronic converters. Power electronic devices inject harmonics into the system resulting in a poor power quality for the consumers. This can be overcome by using harmonic filters. Harmonic filters may be active, passive or hybrid and must be selected appropriately according to the requirement. In addition the starting of directly connected induction generators may cause voltage flicker. The soft-staring mechanisms are sometimes useful in reducing these voltage fluctuations. Wind generators can also introduce periodic fluctuations in the voltages due to tower shadow effect [Fadaeinedjad et al, 2009]. 2.2.4 Interconnection standards and examples of grid codes In order to address protection and safety issues, DG interconnection standards and guidelines have been introduced. The most widely applied standard is the IEEE Std. 1547-2003, "IEEE Standards for Interconnecting Distributed Recourses (DR) with Electric Power Systems 2003" [IEEE, 2003]. Most utilities have adopted the IEEE Std. 1547-2003 in formulating guidelines/ rules for interconnecting DG to their networks. A brief description of the salient features of IEEE Std. 1547-2003 is given below. IEEE Std. 1547-2003 recommends disconnection of distributed energy resources when the voltage and/or frequency at the point of interconnection deviate from their base values due to faults and other disturbances. The allowable deviations are defined in Tables 2 and 3. Also, for an unintentional island in which the DG energizes a portion of the network, the DG DG DG Load PCC PCC PCC PCC: point of common couplingLoad Load A B C G Utility Grid Grid Integration of Renewable Energy Systems 113Voltage magnitude Utility Grid Renewable Energy112 operate at two different frequencies [Jenkins et al, 2000]. Reclosing of the auto recloser while the two systems are out of phase could bring disastrous results. [Borbely & Kreider, 2000]. Further, islanding operation may create an ungrounded system depending on the transformer connection. An unidentified island would be hazardous for the repair crews. Due to the above reasons, islanding is recognized as an unsafe situation and immediate disconnection of the DGs from the grid is recommended upon formation of an island [IEEE, 2003]. 2.2.2 Voltage control Interconnection of DG results in changes in power flows and the voltage profile of the feeder, and generally results in overvoltages under low load or high (DG) production conditions. In weak networks, the DG capacity is generally determined by the voltage limits. Furthermore, the connection status of a DG is not controlled by the utility. Disconnection of a DG during the high load can cause undervoltages, while re-connection of a DG under low load conditions may cause overvoltages. This may lead to poor power quality situation and may result in operation of under/over voltage relays. Selection of appropriate tap settings for the distribution transformers becomes difficult with the increased penetration of DG. This is specially difficult when the DGs are not equally distributed among the feeders supplied by the same transformer [Uchida et al, 2006]. Such a situation is illustrated in Figure 2 where there are two feeders supplied by the same transformer, but DGs are concentrated on only one of them. Generally, when the DGs are connected, the net current flow through the transformer is reduced because the DGs provide the power to the nearby loads. Consequently, the transformer tap needs to be changed to the light load setting. The resulting decline of the sending voltage can cause a voltage violation at the far end of the feeder without DGs as shown in Figure 2. Leaving the transformer tap at the heavy load setting risks overvoltages on the feeder with DGs. Switched capacitors and static VAr compensators can be used to control the feeder voltages, but these solutions are often too costly. Fig. 2. Possible DG interconnection configurations Grid DG DGVoltage magnitude Heavy load settingDistribution distance VmaxLight load settingVmin Feeder with DG Feeder without DGFeeder with DG Feeder without DGVoltage limit violations Another issue that affects the operation of DGs at distribution level is the unbalanced voltage profile. As shown in Figure 3, both the loads as well as DGs can be either three-phase or single-phase. Interconnection of single phase sources will increase the system unbalance. On the other hand, inherently unbalanced distribution systems can pose problems for the three-phase DGs connected to it: the resulting unbalance currents in the DG can cause overheating and frequent shutdowns. Fig. 3. Possible DG interconnection configurations 2.2.3 Harmonics and flicker The quality of power supplied to the customers is one of the major concerns of service providers. The majority of the DGs are interfaced using power electronic converters. Power electronic devices inject harmonics into the system resulting in a poor power quality for the consumers. This can be overcome by using harmonic filters. Harmonic filters may be active, passive or hybrid and must be selected appropriately according to the requirement. In addition the starting of directly connected induction generators may cause voltage flicker. The soft-staring mechanisms are sometimes useful in reducing these voltage fluctuations. Wind generators can also introduce periodic fluctuations in the voltages due to tower shadow effect [Fadaeinedjad et al, 2009]. 2.2.4 Interconnection standards and examples of grid codes In order to address protection and safety issues, DG interconnection standards and guidelines have been introduced. The most widely applied standard is the IEEE Std. 1547-2003, "IEEE Standards for Interconnecting Distributed Recourses (DR) with Electric Power Systems 2003" [IEEE, 2003]. Most utilities have adopted the IEEE Std. 1547-2003 in formulating guidelines/ rules for interconnecting DG to their networks. A brief description of the salient features of IEEE Std. 1547-2003 is given below. IEEE Std. 1547-2003 recommends disconnection of distributed energy resources when the voltage and/or frequency at the point of interconnection deviate from their base values due to faults and other disturbances. The allowable deviations are defined in Tables 2 and 3. Also, for an unintentional island in which the DG energizes a portion of the network, the DG DG DG Load PCC PCC PCC PCC: point of common couplingLoad Load A B C G Utility Grid Grid Integration of Renewable Energy Systems 113Voltage magnitude Utility Grid Renewable Energy114 DG interconnection system must detect the island and disconnect the DG within two seconds of the formation of an island [IEEE, 2003]. Voltage range (% of base voltage a ) Clearing time(s) b V < 50 % 0.16 50 %  V < 88 % 2.00 110 %  V < 120 % 1.00 V  120 % 0.16 aBase voltages are the nominal system voltages stated in ANSI C84.1-1995. bDR ≤ 30 kW, maximum clearing times; DR > 30kW, default clearing times. Table 2. Interconnection System Response to Abnormal Voltages [IEEE, 2003] DR size Frequency range(Hz) Clearing time(s) a  30kW > 60.5 0.16 < 59.3 0.16 > 30kW > 60.5 0.16 < {59.8-57.0} (adjustable set point) 0.16 to 300 (adjustable) < 57.0 0.16 aDR ≤ 30 kW, maximum clearing times; DR > 30kW, default clearing times. Table 3. Interconnection System Response to Abnormal Frequencies [IEEE, 2003] 2.3 Islanding and anti-islanding protection methods As discussed earlier, formation of unintentional power islands is unsafe an undesirable. Therefore, formation of islands must be swiftly detected and the generators energizing the island must be disconnected. This is commonly known as the anti-islanding protection or the loss-of-main protection. There are several methods that provide anti-islanding protection. They can be broadly categorized into three groups: active methods, passive methods and telecommunication based methods. 2.3.1 Passive methods The passive islanding detection methods make decisions based on measured electrical quantities such as voltage and frequency. Voltage relay is one of the simplest passive methods used to detect islands. The reactive power imbalance between production and consumption, which occurs after the loss of mains, leads to a change in the voltage level. Thus, the voltage magnitude measured at the DG can be used to detect the island. Trip signals are generated if the measured voltage shows abnormal variation over a predefined period of time. Normally, voltage relays respond to both under-voltage and over-voltage situations. However, voltage collapse is a slow process and detection of islands using voltage relays could be very slow. Voltage Surge relay (also known as Voltage Vector Shift relay or Voltage Phase Jump relay) is one of the methods used for fast detection of islands. The loss of grid connection results in a sudden change in load and thus, a sudden change in cycle length of the terminal voltage waveform. The cycle duration becomes either shorter or longer, depending on if there is an excess or a deficit of active power in the islanded system. The Vector Shift relay measures the length of each cycle of the voltage waveform. The cycle duration of the waveform is compared with the duration of the previous cycle. If the difference in the cycle duration exceeds the relay setting, a trip signal is issued. Usually these relays are provided with low voltage blocking facility to block the trip signal if the terminal voltage drops below a threshold level to avoid false operations [Freitas et al, 2005]. Frequency relay is another example of passive detection methods. Under-frequency can occur if the grid connection is lost at a situation where the local load exceeds the production of the generator, whereas over-frequency situations can arise if there is a surplus production at the time of grid disconnection. Thus, the frequency relays can take decisions based on the frequency of the voltage at the DG. The over-frequency or under-frequency elements are used to trip the generator from the system. If the real power in the island is almost balanced, the change of frequency of the islanded section will be low, making the relay ineffective. Rate of change of frequency (ROCOF), which is the time derivative of the frequency, is also frequently used to detect power islands. ROCOF relays are popular as they response much faster than the frequency relays. 2.3.2 Active methods In the active detection methods, disturbances are injected into the network and islands are detected based on system responses to the injected disturbance [Freitas et al, 2005]. Reactive error export, fault level monitoring, system impedance monitoring and frequency-drift are some of the active island detection methods. The reactive error export method controls the DG excitation current to generate a known value of reactive current. This cannot be supported when the DG is not connected to the grid. The change in the reactive power measured during the formation of power island is used to detect the situation [Redfern et al, 1993; Kane & Fox, 1997]. The fault level monitoring method uses a point-on-wave switched thyristor in series with a shunt inductor. The thyristor is triggered at a point close to the voltage zero crossing and the current through the shunt inductor is measured. It is possible to estimate the system impedance and the fault level using the measured current pulse. The fact that the fault level drops when the grid connection is lost, is used to detect the islanding. Usually the estimation of fault level is repeated at every half cycle [Redfern et al, 1993, Kane & Fox, 1997]. System impedance monitoring as its name implies, is a method that detects loss of mains (LOM) by actively monitoring the system impedance. A high frequency (HF) source is connected via a coupling capacitor at the point of interconnection (PCC). The capacitor is in series with the equivalent network impedance. With the presence of connection to the utility grid, the equivalent impedance at the PCC is low; therefore the HF-ripple at coupling point is negligible. However, after islanding, the equivalent impedance at the PCC increases dramatically and the divided HF-signal is clearly detectable [Redfern et al, 1993; Kane & Fox, 1997]. The frequency drift LOM detection method is specifically used in inverter interfaced DGs. The output current of the converter is controlled to a frequency which is slightly different to the nominal frequency of the system. Under normal conditions, the terminal frequency is dictated by the powerful bulk supply. If the mains supply is lost, frequency will drift until a certain shutdown level is exceeded [Ropp et al, 1999; Hung et al, 2003]. Grid Integration of Renewable Energy Systems 115Renewable Energy114 DG interconnection system must detect the island and disconnect the DG within two seconds of the formation of an island [IEEE, 2003]. Voltage range (% of base voltage a ) Clearing time(s) b V < 50 % 0.16 50 %  V < 88 % 2.00 110 %  V < 120 % 1.00 V  120 % 0.16 aBase voltages are the nominal system voltages stated in ANSI C84.1-1995. bDR ≤ 30 kW, maximum clearing times; DR > 30kW, default clearing times. Table 2. Interconnection System Response to Abnormal Voltages [IEEE, 2003] DR size Frequency range(Hz) Clearing time(s) a  30kW > 60.5 0.16 < 59.3 0.16 > 30kW > 60.5 0.16 < {59.8-57.0} (adjustable set point) 0.16 to 300 (adjustable) < 57.0 0.16 aDR ≤ 30 kW, maximum clearing times; DR > 30kW, default clearing times. Table 3. Interconnection System Response to Abnormal Frequencies [IEEE, 2003] 2.3 Islanding and anti-islanding protection methods As discussed earlier, formation of unintentional power islands is unsafe an undesirable. Therefore, formation of islands must be swiftly detected and the generators energizing the island must be disconnected. This is commonly known as the anti-islanding protection or the loss-of-main protection. There are several methods that provide anti-islanding protection. They can be broadly categorized into three groups: active methods, passive methods and telecommunication based methods. 2.3.1 Passive methods The passive islanding detection methods make decisions based on measured electrical quantities such as voltage and frequency. Voltage relay is one of the simplest passive methods used to detect islands. The reactive power imbalance between production and consumption, which occurs after the loss of mains, leads to a change in the voltage level. Thus, the voltage magnitude measured at the DG can be used to detect the island. Trip signals are generated if the measured voltage shows abnormal variation over a predefined period of time. Normally, voltage relays respond to both under-voltage and over-voltage situations. However, voltage collapse is a slow process and detection of islands using voltage relays could be very slow. Voltage Surge relay (also known as Voltage Vector Shift relay or Voltage Phase Jump relay) is one of the methods used for fast detection of islands. The loss of grid connection results in a sudden change in load and thus, a sudden change in cycle length of the terminal voltage waveform. The cycle duration becomes either shorter or longer, depending on if there is an excess or a deficit of active power in the islanded system. The Vector Shift relay measures the length of each cycle of the voltage waveform. The cycle duration of the waveform is compared with the duration of the previous cycle. If the difference in the cycle duration exceeds the relay setting, a trip signal is issued. Usually these relays are provided with low voltage blocking facility to block the trip signal if the terminal voltage drops below a threshold level to avoid false operations [Freitas et al, 2005]. Frequency relay is another example of passive detection methods. Under-frequency can occur if the grid connection is lost at a situation where the local load exceeds the production of the generator, whereas over-frequency situations can arise if there is a surplus production at the time of grid disconnection. Thus, the frequency relays can take decisions based on the frequency of the voltage at the DG. The over-frequency or under-frequency elements are used to trip the generator from the system. If the real power in the island is almost balanced, the change of frequency of the islanded section will be low, making the relay ineffective. Rate of change of frequency (ROCOF), which is the time derivative of the frequency, is also frequently used to detect power islands. ROCOF relays are popular as they response much faster than the frequency relays. 2.3.2 Active methods In the active detection methods, disturbances are injected into the network and islands are detected based on system responses to the injected disturbance [Freitas et al, 2005]. Reactive error export, fault level monitoring, system impedance monitoring and frequency-drift are some of the active island detection methods. The reactive error export method controls the DG excitation current to generate a known value of reactive current. This cannot be supported when the DG is not connected to the grid. The change in the reactive power measured during the formation of power island is used to detect the situation [Redfern et al, 1993; Kane & Fox, 1997]. The fault level monitoring method uses a point-on-wave switched thyristor in series with a shunt inductor. The thyristor is triggered at a point close to the voltage zero crossing and the current through the shunt inductor is measured. It is possible to estimate the system impedance and the fault level using the measured current pulse. The fact that the fault level drops when the grid connection is lost, is used to detect the islanding. Usually the estimation of fault level is repeated at every half cycle [Redfern et al, 1993, Kane & Fox, 1997]. System impedance monitoring as its name implies, is a method that detects loss of mains (LOM) by actively monitoring the system impedance. A high frequency (HF) source is connected via a coupling capacitor at the point of interconnection (PCC). The capacitor is in series with the equivalent network impedance. With the presence of connection to the utility grid, the equivalent impedance at the PCC is low; therefore the HF-ripple at coupling point is negligible. However, after islanding, the equivalent impedance at the PCC increases dramatically and the divided HF-signal is clearly detectable [Redfern et al, 1993; Kane & Fox, 1997]. The frequency drift LOM detection method is specifically used in inverter interfaced DGs. The output current of the converter is controlled to a frequency which is slightly different to the nominal frequency of the system. Under normal conditions, the terminal frequency is dictated by the powerful bulk supply. If the mains supply is lost, frequency will drift until a certain shutdown level is exceeded [Ropp et al, 1999; Hung et al, 2003]. Grid Integration of Renewable Energy Systems 115Renewable Energy116 2.3.3 Telecommunication-based Methods The telecommunication-based methods use communicated circuit breaker status signals to alert and trip DGs when islands are formed. Their performance is independent of the type of distributed energy resources involved. Telecommunication based systems can be implemented in conjunction with SCADA systems. In this approach, the states of the circuit breakers in the grid are continuously tracked. The circuit breaker state information can be used to determine whether a part of the system has become an island, using predetermined logic. Transfer tripping schemes can be considered as a decentralized version of the SCADA based system described above. In transfer trip schemes, a logic circuit uses information of circuit breaker states to determine if a part of the grid has been islanded. Comparison of rate of change of frequency (COROCOF) at the substation and the DG location is another telecommunication-based method. The rate of change of frequency is measured at the substation and if it exceeds a certain limit, a block signal is sent to the DG end. If the rate of change of frequency at DG is greater than a set point, and if there is no block signal received from the substation end, the DG will be tripped. Out of the above described methods under/over voltage, under/over frequency, rate of change of frequency (ROCOF) and voltage vector surge (VVS) relays are the most widely used methods for anti-islanding protection. One of the main disadvantages of these methods is the possibility of nuisance tripping of DGs during other system disturbances such as load switching, faults, etc. 2.3.4 Intentional islanding and Microgrids Although islanded operation of DG has recognized as hazardous situation, a generator or cluster of generators operating as an island (microgrid) may be able to supply a part of the local load when grid supply is not available. This type of operation is commonly known as intentional islanding. Numerous technical issues have to be addressed to make intentional islanded operation a reality [Freitas et al, 2005]. The Power balance needs to be maintained between production and consumption. An effective protection coordination must be maintained under both grid connected and islanded modes. In order to reach such a high level of local autonomy, it requires solving advanced control and protection functions. 3. Interconnection studies When interconnecting a renewable energy based generator, the grid interface needs to be designed to meet applicable regulatory requirements set by the local utility. In order to verify the requirements pertaining to protection, control and power quality, detailed simulation studies need to be performed. Time domain simulation using Electromagnetic Transient (EMT) programs is a powerful method that can be used for studies involving controller tuning, protection setting, power quality investigations and system validations. EMT programs typically come with a built-in library that include detailed rotating machine models (synchronous, induction, permanent magnet, etc.), transformers models (including iron core saturation), frequency dependent transmission line and cable models, measurement transformer (CT, VT and CCVT) models from which realistic secondary waveforms can be derived for protection system validation. In addition to electrical components, facilities exist for simulating complex control systems. These computer programs also offer the ability of modeling of power electronic systems in detail. Thus, they can be used for studying all type of DG interfaces:  directly connected synchronous/induction generators,  converter connected DC sources (fuel cell, PV),  doubly fed induction generators (DFIG), and  converter connected variable frequency synchronous generators. Few simulation studies that demonstrate different interconnections interfaces and issues are discussed in the following sections. The simulations were carried out using the industry standard PSCAD/EMTDC software program. 3.1 Soft-starting of fixed speed induction generator wind turbines The first example is a fixed speed wind generator interconnected to a medium voltage (MV) distribution system. The distribution system shown in Figure 4 is adopted from a benchmark MV distribution system proposed by CIGRE for DG integration studies [Rudion, 2005]. The network is normally operated as a radial system, although two disconnect switches (S2 and S3) are available for network reconfiguration. Fig. 4. Medium voltage distribution system Grid Integration of Renewable Energy Systems 117Renewable Energy116 2.3.3 Telecommunication-based Methods The telecommunication-based methods use communicated circuit breaker status signals to alert and trip DGs when islands are formed. Their performance is independent of the type of distributed energy resources involved. Telecommunication based systems can be implemented in conjunction with SCADA systems. In this approach, the states of the circuit breakers in the grid are continuously tracked. The circuit breaker state information can be used to determine whether a part of the system has become an island, using predetermined logic. Transfer tripping schemes can be considered as a decentralized version of the SCADA based system described above. In transfer trip schemes, a logic circuit uses information of circuit breaker states to determine if a part of the grid has been islanded. Comparison of rate of change of frequency (COROCOF) at the substation and the DG location is another telecommunication-based method. The rate of change of frequency is measured at the substation and if it exceeds a certain limit, a block signal is sent to the DG end. If the rate of change of frequency at DG is greater than a set point, and if there is no block signal received from the substation end, the DG will be tripped. Out of the above described methods under/over voltage, under/over frequency, rate of change of frequency (ROCOF) and voltage vector surge (VVS) relays are the most widely used methods for anti-islanding protection. One of the main disadvantages of these methods is the possibility of nuisance tripping of DGs during other system disturbances such as load switching, faults, etc. 2.3.4 Intentional islanding and Microgrids Although islanded operation of DG has recognized as hazardous situation, a generator or cluster of generators operating as an island (microgrid) may be able to supply a part of the local load when grid supply is not available. This type of operation is commonly known as intentional islanding. Numerous technical issues have to be addressed to make intentional islanded operation a reality [Freitas et al, 2005]. The Power balance needs to be maintained between production and consumption. An effective protection coordination must be maintained under both grid connected and islanded modes. In order to reach such a high level of local autonomy, it requires solving advanced control and protection functions. 3. Interconnection studies When interconnecting a renewable energy based generator, the grid interface needs to be designed to meet applicable regulatory requirements set by the local utility. In order to verify the requirements pertaining to protection, control and power quality, detailed simulation studies need to be performed. Time domain simulation using Electromagnetic Transient (EMT) programs is a powerful method that can be used for studies involving controller tuning, protection setting, power quality investigations and system validations. EMT programs typically come with a built-in library that include detailed rotating machine models (synchronous, induction, permanent magnet, etc.), transformers models (including iron core saturation), frequency dependent transmission line and cable models, measurement transformer (CT, VT and CCVT) models from which realistic secondary waveforms can be derived for protection system validation. In addition to electrical components, facilities exist for simulating complex control systems. These computer programs also offer the ability of modeling of power electronic systems in detail. Thus, they can be used for studying all type of DG interfaces:  directly connected synchronous/induction generators,  converter connected DC sources (fuel cell, PV),  doubly fed induction generators (DFIG), and  converter connected variable frequency synchronous generators. Few simulation studies that demonstrate different interconnections interfaces and issues are discussed in the following sections. The simulations were carried out using the industry standard PSCAD/EMTDC software program. 3.1 Soft-starting of fixed speed induction generator wind turbines The first example is a fixed speed wind generator interconnected to a medium voltage (MV) distribution system. The distribution system shown in Figure 4 is adopted from a benchmark MV distribution system proposed by CIGRE for DG integration studies [Rudion, 2005]. The network is normally operated as a radial system, although two disconnect switches (S2 and S3) are available for network reconfiguration. Fig. 4. Medium voltage distribution system Grid Integration of Renewable Energy Systems 117Renewable Energy118 In detailed simulation studies, it is customary to use a reduced model of the network. In the simplified distribution network shown in Figure 5, a section of the network (upstream of bus 8) is replaced by its Thevenin’s equivalent circuit. Parts of the network that are of interest for the study are represented in detail. In this example, it is assumed that bus 11 is supplying a critical load sensitive to power quality problems. Therefore, the feeder section from bus 8 to bus 11 is represented in detail. The direct coupled wind generator is represented using a detailed induction machine model. Although it is not critical for this study, a detailed wind turbine model with pitch control is also used. Fig. 5. Simplified distribution system Time(s) 1.0 2.0 3.0 0.0 3.0 6.0 9.0 12.0 15.0 Voltage(kV)TERMINAL VOLTAGE Fig. 6. Voltage at bus 11 during the wind generator start-up. Time(s) =Time(s) 1.0 2.0 3.0 -2.0 -1.0 0.0 1.0 2.0 Power (MW)REAL POWER GEN-3.0 -2.0 -1.0 0.0 Reac. Power (MVAR)REACTIVE POWER GEN Fig. 7. Wind generator real and reactive power output during the start-up. The usual starting procedure is to let the wind turbine accelerate the generator to a specific predetermined speed (close to the synchronous speed ) and then connect the generator terminal to the grid supply. The machined speeds up, (in motoring mode), drawing both real and reactive power from the grid. This starting procedure has caused a large voltage dip at bus 11 as can be seen in Figure 6. This is a result of the machine initially darwing a significat amount of starting current from the system. The real and reactive power exchange between the induction generator and the grid during this period is shown in Figure 7. One low cost solution to this problem is the use of a thyrister based soft starter [Peters et al, 2006]. The variation of bus 11 voltage with the soft starter is shown in Figure 8. As evident from the simulaion results, the soft starter can significantly reduce the voltage dip. The real and reactive power exchanges with the soft starter is shown in Figure 9. Time(s) 1.0 2.0 3.0 0.0 3.0 6.0 9.0 12.0 15.0 Voltage(kV)TERMINAL VOLTAGE Fig. 8. Voltage at bus 11 during soft starting of the wind generator. i e(s) i e(s) Grid Integration of Renewable Energy Systems 119Voltage(kV)Power (MW)Reac. Power (MVAR)Voltage(kV)Renewable Energy118 In detailed simulation studies, it is customary to use a reduced model of the network. In the simplified distribution network shown in Figure 5, a section of the network (upstream of bus 8) is replaced by its Thevenin’s equivalent circuit. Parts of the network that are of interest for the study are represented in detail. In this example, it is assumed that bus 11 is supplying a critical load sensitive to power quality problems. Therefore, the feeder section from bus 8 to bus 11 is represented in detail. The direct coupled wind generator is represented using a detailed induction machine model. Although it is not critical for this study, a detailed wind turbine model with pitch control is also used. Fig. 5. Simplified distribution system Time(s) 1.0 2.0 3.0 0.0 3.0 6.0 9.0 12.0 15.0 Voltage(kV)TERMINAL VOLTAGE Fig. 6. Voltage at bus 11 during the wind generator start-up. Time(s) =Time(s) 1.0 2.0 3.0 -2.0 -1.0 0.0 1.0 2.0 Power (MW)REAL POWER GEN-3.0 -2.0 -1.0 0.0 Reac. Power (MVAR)REACTIVE POWER GEN Fig. 7. Wind generator real and reactive power output during the start-up. The usual starting procedure is to let the wind turbine accelerate the generator to a specific predetermined speed (close to the synchronous speed ) and then connect the generator terminal to the grid supply. The machined speeds up, (in motoring mode), drawing both real and reactive power from the grid. This starting procedure has caused a large voltage dip at bus 11 as can be seen in Figure 6. This is a result of the machine initially darwing a significat amount of starting current from the system. The real and reactive power exchange between the induction generator and the grid during this period is shown in Figure 7. One low cost solution to this problem is the use of a thyrister based soft starter [Peters et al, 2006]. The variation of bus 11 voltage with the soft starter is shown in Figure 8. As evident from the simulaion results, the soft starter can significantly reduce the voltage dip. The real and reactive power exchanges with the soft starter is shown in Figure 9. Time(s) 1.0 2.0 3.0 0.0 3.0 6.0 9.0 12.0 15.0 Voltage(kV)TERMINAL VOLTAGE Fig. 8. Voltage at bus 11 during soft starting of the wind generator. i e(s) i e(s) Grid Integration of Renewable Energy Systems 119Voltage(kV)Power (MW)Reac. Power (MVAR)Voltage(kV)Renewable Energy120 =Time(s) 1.0 2.0 3.0 -2.0 -1.0 0.0 1.0 2.0 Power (MW)REAL POWER GEN-3.0 -2.0 -1.0 0.0 Reac. Power (MVAR)REACTIVE POWER GEN Fig. 9. Wind generator real and reactive power output with soft starting 3.2 Fault current contribution of different generators The nature of fault current contribution of different DG interface technology can be accurately analysed through detalied electromagnetic transient simulations. The output current produced by a synchronous generator connected to bus 7, during a three-phase short circuit at bus 8 is shown in Figure 10. The output currents of (i) a direct coupled induction generator, (ii) a direct coupled induction generator with external rotor resistance, and (iii) a dobly fed induction generator (DFIG) for the same fault are shown in Figs. 11, 12, and 13 respectively. The lack of sustained fault current in the latter three cases is clearly vissible. It should be noted that with power electronic intrefacing (e.g. DFIG, PV), the shape of the fault current is influenced by the converter portection and control stratergy. Time(s) 10.000 10.050 10.100 10.150 -3.0 -2.0 -1.0 0.0 1.0 2.0 3.0 Current(kA)THREE PHASE CURRENT Fig. 10. Fault current contribution from a syncronous generator i e(s) Time(s) Time ... 10.000 10.050 10.100 10.150 -3.0 -2.0 -1.0 0.0 1.0 2.0 3.0 Current(kA)THREE PHASE CURRENT Fig. 11. Fault current contribution from an induction generator Time ... 10.000 10.050 10.100 10.150 -3.0 -2.0 -1.0 0.0 1.0 2.0 3.0 Current(kA)THREE PHASE CURRENT Fig. 12. Fault current contribution from an induction generator with external rotor resistance. Note the faster decay than in the normal induction generator case in Figure 11. Time(s) 10.000 10.050 10.100 10.150 -3.0 -2.0 -1.0 0.0 1.0 2.0 3.0 Current(kA)THREE PHASE CURRENT Fig. 13. Fault current contribution from a doubly fed induction generator (crow bar operation is not simulated). Time(s) Time(s) Time(s) Grid Integration of Renewable Energy Systems 121Power (MW)Reac. Power (MVAR)Current(kA)Current(kA)Current(kA)Current(kA)Renewable Energy120 =Time(s) 1.0 2.0 3.0 -2.0 -1.0 0.0 1.0 2.0 Power (MW)REAL POWER GEN-3.0 -2.0 -1.0 0.0 Reac. Power (MVAR)REACTIVE POWER GEN Fig. 9. Wind generator real and reactive power output with soft starting 3.2 Fault current contribution of different generators The nature of fault current contribution of different DG interface technology can be accurately analysed through detalied electromagnetic transient simulations. The output current produced by a synchronous generator connected to bus 7, during a three-phase short circuit at bus 8 is shown in Figure 10. The output currents of (i) a direct coupled induction generator, (ii) a direct coupled induction generator with external rotor resistance, and (iii) a dobly fed induction generator (DFIG) for the same fault are shown in Figs. 11, 12, and 13 respectively. The lack of sustained fault current in the latter three cases is clearly vissible. It should be noted that with power electronic intrefacing (e.g. DFIG, PV), the shape of the fault current is influenced by the converter portection and control stratergy. Time(s) 10.000 10.050 10.100 10.150 -3.0 -2.0 -1.0 0.0 1.0 2.0 3.0 Current(kA)THREE PHASE CURRENT Fig. 10. Fault current contribution from a syncronous generator i e(s) Time(s) Time ... 10.000 10.050 10.100 10.150 -3.0 -2.0 -1.0 0.0 1.0 2.0 3.0 Current(kA)THREE PHASE CURRENT Fig. 11. Fault current contribution from an induction generator Time ... 10.000 10.050 10.100 10.150 -3.0 -2.0 -1.0 0.0 1.0 2.0 3.0 Current(kA)THREE PHASE CURRENT Fig. 12. Fault current contribution from an induction generator with external rotor resistance. Note the faster decay than in the normal induction generator case in Figure 11. Time(s) 10.000 10.050 10.100 10.150 -3.0 -2.0 -1.0 0.0 1.0 2.0 3.0 Current(kA)THREE PHASE CURRENT Fig. 13. Fault current contribution from a doubly fed induction generator (crow bar operation is not simulated). Time(s) Time(s) Time(s) Grid Integration of Renewable Energy Systems 121Power (MW)Reac. Power (MVAR)Current(kA)Current(kA)Current(kA)Current(kA)Renewable Energy122 3.3 PV grid integration Interconnection of a 250 kW photovoltaic (PV) power system in place of the wind turbine is considered in this example. Detailed models of the PV array, power converter and its controller with maximum power point tracking (MPPT) are used in the simulation. The system behaviour under rapidly changing solar radiation intensity is examined. The schematic of the PV system in PSCAD/EMTDC is shown in Figure 14. Figure 15 shows the variation of the solar radiation intensity. The voltage at bus 14 under the simulated solar radiation variation is shown in Figure 16. The real and reactive power output of the PV inverter is shown in Figure 17. The inverter output current shown in Figure 18 is highly distorted, and its harmonic spectrum is shown in Figure 19. This distorted current causes voltage distortion. The harmonics spectrum of the voltage at bus 14 is shown in Figure 20. If the harmonic distortion is unacceptable, it is necessary to install a filter at the PCC of the PV power system. gt1gt2gt3gt4gt5gt61EabEbcIdc352 6 41352 6 4Ia_griddcVltg#1 #2P = 0.2036Q = -0.01998V = 11VA2e-4800005100005T1VpvIcon0.01 [H]V+GTcellV-SolRadCellTempIpv Fig. 14. Schematic of the grid connected PV system simulation (for the simplicity, the ac system is shown as an equivalent source). Time(s) 2.5 5.0 7.5 10.0 12.5 15.0 17.5 20.0 0.4k0.5k0.6k0.7k0.8k0.9k1.0kRadiation(W/m*m)RADIATION Fig. 15. Variation of the solar radiatin intensity i e(s) 2.5 5.0 7.5 10.0 12.5 15.0 17.5 20.0 0.0 3.0 6.0 9.0 12.0 15.0 Voltage (kV)Vrms Fig. 16. Variation of the voltage at bus 7 Time(s) 2.5 5.0 7.5 10.0 12.5 15.0 17.5 20.0 0.050 0.100 0.150 0.200 0.250 0.300 Power (MW)P-0.100 -0.080 -0.060 -0.040 -0.020 0.000 Reactive Power (MVAr)Q Fig. 17. Real and reactive power output of the inverter Time(s) 7.490 7.500 7.510 7.520 7.530 7.540 7.550 7.560 7.570 -0.0100 -0.0050 0.0000 0.0050 0.0100 Current (kA)THREE PHASE CURRENT Fig. 18. Inverter output current 00.0050.010.0150.022 3 4 5 6 7 8 9 10 11 12 13 14 15Harmonic numberCurrent(kA) Fig. 19. Harmonic spectrum of the inverter current (Fundamental 0.28 kA) Time(s) Ti ) i e( Grid Integration of Renewable Energy Systems 12380000100005Radiation(W/m*m)Voltage (kV)Power (MW)Reactive Power (MVAr)Current (kA)Current(kA)Renewable Energy122 3.3 PV grid integration Interconnection of a 250 kW photovoltaic (PV) power system in place of the wind turbine is considered in this example. Detailed models of the PV array, power converter and its controller with maximum power point tracking (MPPT) are used in the simulation. The system behaviour under rapidly changing solar radiation intensity is examined. The schematic of the PV system in PSCAD/EMTDC is shown in Figure 14. Figure 15 shows the variation of the solar radiation intensity. The voltage at bus 14 under the simulated solar radiation variation is shown in Figure 16. The real and reactive power output of the PV inverter is shown in Figure 17. The inverter output current shown in Figure 18 is highly distorted, and its harmonic spectrum is shown in Figure 19. This distorted current causes voltage distortion. The harmonics spectrum of the voltage at bus 14 is shown in Figure 20. If the harmonic distortion is unacceptable, it is necessary to install a filter at the PCC of the PV power system. gt1gt2gt3gt4gt5gt61EabEbcIdc352 6 41352 6 4Ia_griddcVltg#1 #2P = 0.2036Q = -0.01998V = 11VA2e-4800005100005T1VpvIcon0.01 [H]V+GTcellV-SolRadCellTempIpv Fig. 14. Schematic of the grid connected PV system simulation (for the simplicity, the ac system is shown as an equivalent source). Time(s) 2.5 5.0 7.5 10.0 12.5 15.0 17.5 20.0 0.4k0.5k0.6k0.7k0.8k0.9k1.0kRadiation(W/m*m)RADIATION Fig. 15. Variation of the solar radiatin intensity i e(s) 2.5 5.0 7.5 10.0 12.5 15.0 17.5 20.0 0.0 3.0 6.0 9.0 12.0 15.0 Voltage (kV)Vrms Fig. 16. Variation of the voltage at bus 7 Time(s) 2.5 5.0 7.5 10.0 12.5 15.0 17.5 20.0 0.050 0.100 0.150 0.200 0.250 0.300 Power (MW)P-0.100 -0.080 -0.060 -0.040 -0.020 0.000 Reactive Power (MVAr)Q Fig. 17. Real and reactive power output of the inverter Time(s) 7.490 7.500 7.510 7.520 7.530 7.540 7.550 7.560 7.570 -0.0100 -0.0050 0.0000 0.0050 0.0100 Current (kA)THREE PHASE CURRENT Fig. 18. Inverter output current 00.0050.010.0150.022 3 4 5 6 7 8 9 10 11 12 13 14 15Harmonic numberCurrent(kA) Fig. 19. Harmonic spectrum of the inverter current (Fundamental 0.28 kA) Time(s) Ti ) i e( Grid Integration of Renewable Energy Systems 12380000100005Radiation(W/m*m)Voltage (kV)Power (MW)Reactive Power (MVAr)Current (kA)Current(kA)Renewable Energy124 00.0040.0080.0122 3 4 5 6 7 8 9 10 11 12 13 14 15Harmonic numberVoltage(kV) Fig. 20. Harmonic spectrum of the voltage at bus 7 (Fundamental 7.43kV) 4. Integration of large scale renewable energy generation Large scale renewable energy based generation such as conventional hydro, steam turbines operated with biomass or geothermal energy use conventional technology based on synchronous generators. Utilities are well experienced with this established technology. However, interconnection of large scale wind farms poses entirely different set of challenges, mainly due to the intermittent nature of the wind and the desirability for variable speed operation of the generators. 4.1 Issues related to grid integration of large wind farms Recent rapid growth of wind energy generation has resulted in the development of large wind farms with capacities in excess of 100 MW. Such large scale wind farms are generally interconnected to the transmission grid. In some countries like Germany and Spain, the installed wind generation capacity has reached several thousand megawatts and the installed wind capacity is continuously increasing. However, power output of the wind generators cannot be dispatched. Despite the advances in short-term (and long-term) wind forecasting technology, accurate forecasting of wind speed is difficult. Therefore, a significant spinning reserve and stand by capacity is required to cover the fluctuations in wind farm production. Although the distribution of wind farms over a wide geographic area provides some cushioning effect, the possibility of simultaneous low production from all wind farms cannot be entirely ruled out. During a disturbance such as a short duration voltage dip caused by a short circuit in the high voltage transmission network, several thousand megawatts of wind generation can get disconnected depriving real power to the grid. Such a condition can result in serious stability problems within the power system. Thus, new connection requirements for large wind farms have been introduced by many system operators. These new connection regulations such as fault ride through capabilities compel wind turbines to have a supporting effect on grid operation. A typical fault ride through capability requirement, expected from large, modern wind farms is shown in Figure 21. It specifies the time duration that the wind farm is required to stay connected, for different levels of voltage depressions. This is an attempt to ensure that the wind farm is capable and ready to supply power to the system immediately after clearing the fault to maintain system stability. Fig. 21. Typical fault ride through requirements. The actual fault ride through curve can vary depending on the transmission operator. The ‘strict’ fault ride through standards presents serious challenges to the wind turbine system designer.  During the fault, the power transfer capability of the wind turbine generator (WTG) is diminished. This will result in mechanical over speeding of the machine and the turbine blades. This will place mechanical stress on blades and the other rotating parts.  During the fault recovery period, depending on the generator technology employed, the WTG may draw increased amounts of reactive power from the system. This will result in a poor fault recovery response.  The nature and the magnitude of the fault current contribution from the wind farm depend on the WTG technology. A good understanding of this is necessary to design equipment ratings and for protection and control settings.  Turbine blade pitch control must be capable of regulating the input mechanical power. During faults, fast pitch control of blades will limit the mechanical over speeding of rotating elements.  Application of appropriate FACTS based reactive power compensation devices can help improve the overall performance of the wind farm. The requirement of reactive power for voltage support is one of the key issues related to wind power generation. Most wind generation systems use induction generators. Unlike synchronous machines, induction machines do not have reactive power control capability. Interconnection regulations require the wind farms to maintain the power factor within a narrow range, usually within 0.95. If the wind generators are not capable of meeting this requirement, dynamic reactive support must be provided. Usually the reactive power support is provided using a combination of a capacitor bank and static var compensators (SVCs)/STATCOM. The sudden disconnection of the wind turbine generators while the capacitor bank is still connected can result in overvoltage at the generator terminals. This is a major protection problem related to the large wind farms. The type of reactive power compensation method and devices used at a wind farm also plays an important role in the performance of the wind farm during the period immediately following the clearance of a system fault. Reactive power capability of a capacitor bank (and V/Vn Time (s) 100% 60% 80% 40% 20% 0 1.0 2.0 3.0 4.0WTG disconnection not allowed WTG disconnection allowed Grid Integration of Renewable Energy Systems 125Voltage(kV)Renewable Energy124 00.0040.0080.0122 3 4 5 6 7 8 9 10 11 12 13 14 15Harmonic numberVoltage(kV) Fig. 20. Harmonic spectrum of the voltage at bus 7 (Fundamental 7.43kV) 4. Integration of large scale renewable energy generation Large scale renewable energy based generation such as conventional hydro, steam turbines operated with biomass or geothermal energy use conventional technology based on synchronous generators. Utilities are well experienced with this established technology. However, interconnection of large scale wind farms poses entirely different set of challenges, mainly due to the intermittent nature of the wind and the desirability for variable speed operation of the generators. 4.1 Issues related to grid integration of large wind farms Recent rapid growth of wind energy generation has resulted in the development of large wind farms with capacities in excess of 100 MW. Such large scale wind farms are generally interconnected to the transmission grid. In some countries like Germany and Spain, the installed wind generation capacity has reached several thousand megawatts and the installed wind capacity is continuously increasing. However, power output of the wind generators cannot be dispatched. Despite the advances in short-term (and long-term) wind forecasting technology, accurate forecasting of wind speed is difficult. Therefore, a significant spinning reserve and stand by capacity is required to cover the fluctuations in wind farm production. Although the distribution of wind farms over a wide geographic area provides some cushioning effect, the possibility of simultaneous low production from all wind farms cannot be entirely ruled out. During a disturbance such as a short duration voltage dip caused by a short circuit in the high voltage transmission network, several thousand megawatts of wind generation can get disconnected depriving real power to the grid. Such a condition can result in serious stability problems within the power system. Thus, new connection requirements for large wind farms have been introduced by many system operators. These new connection regulations such as fault ride through capabilities compel wind turbines to have a supporting effect on grid operation. A typical fault ride through capability requirement, expected from large, modern wind farms is shown in Figure 21. It specifies the time duration that the wind farm is required to stay connected, for different levels of voltage depressions. This is an attempt to ensure that the wind farm is capable and ready to supply power to the system immediately after clearing the fault to maintain system stability. Fig. 21. Typical fault ride through requirements. The actual fault ride through curve can vary depending on the transmission operator. The ‘strict’ fault ride through standards presents serious challenges to the wind turbine system designer.  During the fault, the power transfer capability of the wind turbine generator (WTG) is diminished. This will result in mechanical over speeding of the machine and the turbine blades. This will place mechanical stress on blades and the other rotating parts.  During the fault recovery period, depending on the generator technology employed, the WTG may draw increased amounts of reactive power from the system. This will result in a poor fault recovery response.  The nature and the magnitude of the fault current contribution from the wind farm depend on the WTG technology. A good understanding of this is necessary to design equipment ratings and for protection and control settings.  Turbine blade pitch control must be capable of regulating the input mechanical power. During faults, fast pitch control of blades will limit the mechanical over speeding of rotating elements.  Application of appropriate FACTS based reactive power compensation devices can help improve the overall performance of the wind farm. The requirement of reactive power for voltage support is one of the key issues related to wind power generation. Most wind generation systems use induction generators. Unlike synchronous machines, induction machines do not have reactive power control capability. Interconnection regulations require the wind farms to maintain the power factor within a narrow range, usually within 0.95. If the wind generators are not capable of meeting this requirement, dynamic reactive support must be provided. Usually the reactive power support is provided using a combination of a capacitor bank and static var compensators (SVCs)/STATCOM. The sudden disconnection of the wind turbine generators while the capacitor bank is still connected can result in overvoltage at the generator terminals. This is a major protection problem related to the large wind farms. The type of reactive power compensation method and devices used at a wind farm also plays an important role in the performance of the wind farm during the period immediately following the clearance of a system fault. Reactive power capability of a capacitor bank (and V/Vn Time (s) 100% 60% 80% 40% 20% 0 1.0 2.0 3.0 4.0WTG disconnection not allowed WTG disconnection allowed Grid Integration of Renewable Energy Systems 125Voltage(kV)Renewable Energy126 SVC’s) is directly related to the ‘available’ terminal voltage. Due to this reason, capacitor banks have a degraded reactive power capability during fault recovery when in fact, their reactive power capability is most needed. FACTS based voltage source converters (e.g STATCOM) does not have this inherent drawback and is generally the superior option, purely from a reactive power capability point of view. The selection of the reactive power compensation scheme and its control strategy must be carefully evaluated. Synchronous condensers are another reactive power compensation option for large wind farms. They have a slower response compared to a STATCOM, but do provide inertia to the system. This may have advantages, especially when interconnecting to weak ac systems with dynamic stability concerns. Other possible challenges that require detailed investigation may include:  Turbine power electronic design and controller optimization  The design of filters and protection systems  Interactions between numerous turbines  Problems of wind farms connected into series compensated systems  Power quality issues including voltage flicker  Starting and synchronizing of wind farms to the grid.  Sub synchronous resonance issues due to interaction of the electric network and the complex shaft/gear system of the wind turbine. Detailed studies, based on electromagnetic transient programs is essential to identify possible interconnection issues and to find optimum solutions. 4.2 Modeling and simulation requirements 4.2.1 Turbine characteristics The power transferred by the turbine blades is given by (1) 321 ),( wprmech VCAp   (1) wrrVr (2) where Cp is the power coefficient, β is the pitch angle, λ is the tip speed ratio, Vw is the wind speed, Ωr is the rotor speed (on the low-speed side of the gearbox), rr is the rotor-plane radius, ρ is the air density and Ar is the area swept by the rotor. The power coefficient, Cp, represents the aerodynamic characteristics of the blade. Note that in Figure 22, Cp is plotted as a function of the ‘tip speed ratio’. Tip speed ratio is a measure of the relative speed of wind compared to the rotational speed of the blades. The Cp characteristics can be altered through blade pitch angle () control, thereby controlling the amount of mechanical power available to the generator for power conversion. The exact characteristic of Cp depends on the particular design of the blade. The function Cp(λ,β) can be approximated using analytical equations or determined by tests. Fig. 22. Typical variation of Cp with wind speed 4.2.2 Turbine pitch control The control system that orders the pitch angle (β) and the response of the actuators and other mechanical parts may need to be adequately represented in the simulation study. It should be noted that due to the large inertias involved (turbine blades), the overall response of pitch control is ‘slow’ and is in the order of a few seconds. 4.2.3 Wind generator technology Most wind turbine generators at present are based on asynchronous generators. It is advantageous to operate these machines at variable speeds during low wind speed conditions. To produce constant voltage and frequency from these asynchronous generators operating at variable speed, they are often interfaced with the grid using various setups of power electronics and controls. The power electronic interface also allows for other desirable features. The controls are sophisticated and complex. For example they possess the capability of controlling the voltage, real and reactive power output of the machine within some design range, while meeting other constraints such as limiting the positive power ramp rate of the machine. Controller design and settings must ensure low-voltage ride-through compliance. In the future, it is expected that wind turbines would be able to provide more functionalities such as advanced governor functions and controlled ramp down during high wind speed events. There are several wind generator technologies with the more common WTG schemes being,  Induction machines directly connected to the grid with external reactive power compensation  Wound rotor induction machines with an external variable resistor in the rotor circuit to control slip (speed)  Doubly-fed induction machines, which employ an ac-dc-ac converter between the stator and rotor circuits  Conventional or permanent magnet synchronous machines connected to the grid through a fully-rated ac-dc-ac converter. Detailed, high band width models of machines that does not neglect the stator flux transients are required for detailed analysis of such events as  starting and synchronization wind generators, Blade Efficiency, CpTip speed ratio,  50% 30% 40% 20% 10% 0 5.0 10.0 15.0 20.0=2o =4o =6o =8o =10o=15o=25o=pitch angle Grid Integration of Renewable Energy Systems 127Renewable Energy126 SVC’s) is directly related to the ‘available’ terminal voltage. Due to this reason, capacitor banks have a degraded reactive power capability during fault recovery when in fact, their reactive power capability is most needed. FACTS based voltage source converters (e.g STATCOM) does not have this inherent drawback and is generally the superior option, purely from a reactive power capability point of view. The selection of the reactive power compensation scheme and its control strategy must be carefully evaluated. Synchronous condensers are another reactive power compensation option for large wind farms. They have a slower response compared to a STATCOM, but do provide inertia to the system. This may have advantages, especially when interconnecting to weak ac systems with dynamic stability concerns. Other possible challenges that require detailed investigation may include:  Turbine power electronic design and controller optimization  The design of filters and protection systems  Interactions between numerous turbines  Problems of wind farms connected into series compensated systems  Power quality issues including voltage flicker  Starting and synchronizing of wind farms to the grid.  Sub synchronous resonance issues due to interaction of the electric network and the complex shaft/gear system of the wind turbine. Detailed studies, based on electromagnetic transient programs is essential to identify possible interconnection issues and to find optimum solutions. 4.2 Modeling and simulation requirements 4.2.1 Turbine characteristics The power transferred by the turbine blades is given by (1) 321 ),( wprmech VCAp   (1) wrrVr (2) where Cp is the power coefficient, β is the pitch angle, λ is the tip speed ratio, Vw is the wind speed, Ωr is the rotor speed (on the low-speed side of the gearbox), rr is the rotor-plane radius, ρ is the air density and Ar is the area swept by the rotor. The power coefficient, Cp, represents the aerodynamic characteristics of the blade. Note that in Figure 22, Cp is plotted as a function of the ‘tip speed ratio’. Tip speed ratio is a measure of the relative speed of wind compared to the rotational speed of the blades. The Cp characteristics can be altered through blade pitch angle () control, thereby controlling the amount of mechanical power available to the generator for power conversion. The exact characteristic of Cp depends on the particular design of the blade. The function Cp(λ,β) can be approximated using analytical equations or determined by tests. Fig. 22. Typical variation of Cp with wind speed 4.2.2 Turbine pitch control The control system that orders the pitch angle (β) and the response of the actuators and other mechanical parts may need to be adequately represented in the simulation study. It should be noted that due to the large inertias involved (turbine blades), the overall response of pitch control is ‘slow’ and is in the order of a few seconds. 4.2.3 Wind generator technology Most wind turbine generators at present are based on asynchronous generators. It is advantageous to operate these machines at variable speeds during low wind speed conditions. To produce constant voltage and frequency from these asynchronous generators operating at variable speed, they are often interfaced with the grid using various setups of power electronics and controls. The power electronic interface also allows for other desirable features. The controls are sophisticated and complex. For example they possess the capability of controlling the voltage, real and reactive power output of the machine within some design range, while meeting other constraints such as limiting the positive power ramp rate of the machine. Controller design and settings must ensure low-voltage ride-through compliance. In the future, it is expected that wind turbines would be able to provide more functionalities such as advanced governor functions and controlled ramp down during high wind speed events. There are several wind generator technologies with the more common WTG schemes being,  Induction machines directly connected to the grid with external reactive power compensation  Wound rotor induction machines with an external variable resistor in the rotor circuit to control slip (speed)  Doubly-fed induction machines, which employ an ac-dc-ac converter between the stator and rotor circuits  Conventional or permanent magnet synchronous machines connected to the grid through a fully-rated ac-dc-ac converter. Detailed, high band width models of machines that does not neglect the stator flux transients are required for detailed analysis of such events as  starting and synchronization wind generators, Blade Efficiency, CpTip speed ratio,  50% 30% 40% 20% 10% 0 5.0 10.0 15.0 20.0=2o =4o =6o =8o =10o=15o=25o=pitch angle Grid Integration of Renewable Energy Systems 127Renewable Energy128  accurate fault ride through analysis  DFIG control and protection (crow bar etc.)  VSC based back to back ac-dc-ac converter design including dc line/cable resonance issues. 4.2.4 Power system model Electromagnetic transient simulations allow for accurate representation of power system equipment. This may include nonlinear core models of transformers and reactors, induction motor and other dynamic loads, overhead lines and cables considering inter-conductor mutual effects, accurate representation of power electronic switching devices, FACTS devices and also control systems. In general, grid integration requires understanding the needs of the power system, knowledge of machine capabilities, and application knowledge of the power converter and control capability to put the entire picture together. Thus a typical study needs modelling of the grid, wind resource, wind turbine, generator, power electronics, controllers and protection. 4.3 Simulation example In this example, the application of an SVC to enhance voltage control and system stability is demonstrated. The single line diagram of the high voltage transmission system used for the study is shown in Figure23 [Jiang et al, 2005]. Fig. 23. High voltage transmission system model. A wind farm is connected to Bus 12. For this study, all generators in the wind farm are represented by a single equivalent generator. In order to provide voltage support, an SVC is connected at bus 6. A fault is applied on the 230 kV transmission line connecting Bus 1 and 1 G 1G 2G 4G 39 2 3 4 5 6 7 8 1 0 1 1 1 2 SV Wind Farm Bus 6. The fault is cleared by opening the line breakers after 9 cycles. The response of the wind generators with and without the SVC in operation is shown in Figs. 24 and 25. It is clear that the wind farm is unstable for this fault if not for the voltage support from the SVC. 3.0 4.0 5.0 6.0 7.0 0.80 0.90 1.00 1.10 Speed (pu)w Fig. 24. With the SVC in operation, the generator speed reaches a stable steady state upon the removal of the fault 3.0 4.0 5.0 6.0 7.0 0.80 0.90 1.00 1.10 Speed (pu)w Fig. 25. Without the SVC in operation, the wind generator is unstable. 5. Conclusion Various issues and challenges that need to be addressed in grid integration of renewable energy systems are discussed. The major concerns in interconnection of the small scale renewable energy generation at the distribution level are related to protection, voltage control and power quality. The reserve requirements, reactive power requirements and the grid support during the disturbances are among the major issues to be considered in grid interconnection of the large wind farms. The importance of the simulation tools in grid integration studies is demonstrated through several case studies. Time(s) Time(s) Grid Integration of Renewable Energy Systems 129Speed (pu)Speed (pu)Renewable Energy128  accurate fault ride through analysis  DFIG control and protection (crow bar etc.)  VSC based back to back ac-dc-ac converter design including dc line/cable resonance issues. 4.2.4 Power system model Electromagnetic transient simulations allow for accurate representation of power system equipment. This may include nonlinear core models of transformers and reactors, induction motor and other dynamic loads, overhead lines and cables considering inter-conductor mutual effects, accurate representation of power electronic switching devices, FACTS devices and also control systems. In general, grid integration requires understanding the needs of the power system, knowledge of machine capabilities, and application knowledge of the power converter and control capability to put the entire picture together. Thus a typical study needs modelling of the grid, wind resource, wind turbine, generator, power electronics, controllers and protection. 4.3 Simulation example In this example, the application of an SVC to enhance voltage control and system stability is demonstrated. The single line diagram of the high voltage transmission system used for the study is shown in Figure23 [Jiang et al, 2005]. Fig. 23. High voltage transmission system model. A wind farm is connected to Bus 12. For this study, all generators in the wind farm are represented by a single equivalent generator. In order to provide voltage support, an SVC is connected at bus 6. A fault is applied on the 230 kV transmission line connecting Bus 1 and 1 G 1G 2G 4G 39 2 3 4 5 6 7 8 1 0 1 1 1 2 SV Wind Farm Bus 6. The fault is cleared by opening the line breakers after 9 cycles. The response of the wind generators with and without the SVC in operation is shown in Figs. 24 and 25. It is clear that the wind farm is unstable for this fault if not for the voltage support from the SVC. 3.0 4.0 5.0 6.0 7.0 0.80 0.90 1.00 1.10 Speed (pu)w Fig. 24. With the SVC in operation, the generator speed reaches a stable steady state upon the removal of the fault 3.0 4.0 5.0 6.0 7.0 0.80 0.90 1.00 1.10 Speed (pu)w Fig. 25. Without the SVC in operation, the wind generator is unstable. 5. Conclusion Various issues and challenges that need to be addressed in grid integration of renewable energy systems are discussed. The major concerns in interconnection of the small scale renewable energy generation at the distribution level are related to protection, voltage control and power quality. The reserve requirements, reactive power requirements and the grid support during the disturbances are among the major issues to be considered in grid interconnection of the large wind farms. The importance of the simulation tools in grid integration studies is demonstrated through several case studies. Time(s) Time(s) Grid Integration of Renewable Energy Systems 129Speed (pu)Speed (pu)Renewable Energy130 6. References Alderfer, R. B.; Eldridge, M. M. and Starrs, T. J. (2000), Making Connections: Case Studies of Interconnection Barriers and their Impact on Distributed Power Projects, Report No. NREL/SR-200-28053, National Renewable Energy Laboratory, Golden, Colorado, USA, May 2000. Barker, P. & de Mello, R.W. (2000), Determining the Impact of Distributed Generation on Power Systems: Part 1- Radial Power Systems, Proc. IEEE Power Eng. Soc. Summer Meeting, pp.1645-1658, July 2000. Borbely, A.M. & Kreider, J.F. (2000), Distributed Generation, The Power Paradigm for the New Millennium, LLC: CRC Press, 2000. Dugan, B. & McDermott, T. (2002) Distributed Generation: Operating Conflicts for Distributed Generation Interconnected with Utility Distribution System, IEEE Industry Applications Magazine, March/April 2002, pp: 19-25. Fadaeinedjad, R.; Moschopoulos, G. & Moallem, M. (2009), The Impact of Tower Shadow, Yaw Error, and Wind Shears on Power Quality in a Wind–Diesel System, IEEE Transactions on Energy Conversion, Vol. 24, No. 1, March 2009, pp: 102-111. Freitas, W.; Huang, Z. & Xu W.(2005), A Practical Method for Assessing the Effectiveness of Vector Surge Relays for Distributed Generation Applications, IEEE Transactions On Power Delivery, Vol. 20, No. 1, January 2005, pp: 57-63. Freitas, W.; Wilsun Xu; Affonso, C.M. & Zhenyu, Huang (2005), Comparative analysis between ROCOF and vector surge relays for distributed generation applications, IEEE Transactions on Power Delivery, April 2005. pp: 1315 – 1324. Hung, Guo-Kiang.; Chang, Chih-Chang & Chen, Chern-Lin (2003), Automatic Phase-Shift Method for Islanding Detection of Grid-Connected Photovoltaic Inverters, IEEE Transactions On Energy Conversion, Vol. 18, No. 1, March 2003, pp: 169-173. IEEE (2003), IEEE Standard for Interconnecting Distributed Resources with Electric Power Systems, Std. 1547-2003. IEEE (2004), Impact of Distributed Resources on Distribution Relay Protection, Special report of the IEEE, Power System Relay Committee Working Group D3, August 2004. Jenkins, N.; Allan, R.; Crossley, P.; Kirschen, D. & Strbac, G. (2000), Embedded Generation, IEE Power and Energy Series 13, 2000. Jiang, S.; Annakkage, U.D. & Gole, A. M. (2005) , A platform for validation of FACTS models, IEEE Trans. Power Delivery, Vol. 20, No.4, October 2005. Kane, P.O. & Fox, B. (1997), Loss of Mains Detection for Embedded Generation by System Impedance Monitoring Developments in Power System Protection, IEE Conference Publication No. 434, 25-27th March 1997, pp: 95-98. Lopes, J.A.P. (2002), Integration of dispersed generation on distribution networks impact studies, IEEE Power Engineering Society Winter Meeting, Vol. 1, pages 323328, 2002. Mozina, C.J. (2001), Interconnection protection of IPP generators at commercial/industrial facilities, IEEE Transactions on Industry Applications, pp 681-688, 2001. Peters, R.R.; Muthumuni, D.; Bartel, T.; Salehfar, H. & Mann, M. (2006), Dynamic model development of a fixed speed stall control wind turbine at start-up, IEEE Power Engineering Society General Meeting, pp: June 2006, pp: 1-7. Redfern, M. A.; Usta, O. & Fielding, G. (1993), Protection Against Loss Of Utility Grid Supply For A Dispersed Storage And Generation Unit, IEEE Transactions on Power Delivery, Vol. 8, No. 3, July 1993, pp: 948-954. Ropp, M. E. ; Begovic, M. & Rohatgi, A. (1999), Analysis and Performance Assessment of the Active Frequency Drift Method of Islanding Prevention, IEEE Transactions on Energy Conversion, Vol. 14, No. 3, September 1999, pp: 810–816. Rudion, K.; Orths, A.; Styczynski, Z. & Strunz K.,(2005), Design of benchmark of medium voltage distribution network for investigation of DG integration, Proc. Conference on the Innovations, Commercial Applications and Regulatory Frameworks of Distributed Generation. November 30, December 1-2, 2005, Boston, USA. Uchida, A.; Watanabe,S. & Iwamoto, S. (2006), A Voltage Control Strategy for Distribution Networks with Dispersed Generations, Proceeding of IEEE – PES General Meeting, pp. 1-6, Tampa, Florida, USA, June 2007. Grid Integration of Renewable Energy Systems 131Renewable Energy130 6. References Alderfer, R. B.; Eldridge, M. M. and Starrs, T. J. (2000), Making Connections: Case Studies of Interconnection Barriers and their Impact on Distributed Power Projects, Report No. NREL/SR-200-28053, National Renewable Energy Laboratory, Golden, Colorado, USA, May 2000. Barker, P. & de Mello, R.W. (2000), Determining the Impact of Distributed Generation on Power Systems: Part 1- Radial Power Systems, Proc. IEEE Power Eng. Soc. Summer Meeting, pp.1645-1658, July 2000. Borbely, A.M. & Kreider, J.F. (2000), Distributed Generation, The Power Paradigm for the New Millennium, LLC: CRC Press, 2000. Dugan, B. & McDermott, T. (2002) Distributed Generation: Operating Conflicts for Distributed Generation Interconnected with Utility Distribution System, IEEE Industry Applications Magazine, March/April 2002, pp: 19-25. Fadaeinedjad, R.; Moschopoulos, G. & Moallem, M. (2009), The Impact of Tower Shadow, Yaw Error, and Wind Shears on Power Quality in a Wind–Diesel System, IEEE Transactions on Energy Conversion, Vol. 24, No. 1, March 2009, pp: 102-111. Freitas, W.; Huang, Z. & Xu W.(2005), A Practical Method for Assessing the Effectiveness of Vector Surge Relays for Distributed Generation Applications, IEEE Transactions On Power Delivery, Vol. 20, No. 1, January 2005, pp: 57-63. Freitas, W.; Wilsun Xu; Affonso, C.M. & Zhenyu, Huang (2005), Comparative analysis between ROCOF and vector surge relays for distributed generation applications, IEEE Transactions on Power Delivery, April 2005. pp: 1315 – 1324. Hung, Guo-Kiang.; Chang, Chih-Chang & Chen, Chern-Lin (2003), Automatic Phase-Shift Method for Islanding Detection of Grid-Connected Photovoltaic Inverters, IEEE Transactions On Energy Conversion, Vol. 18, No. 1, March 2003, pp: 169-173. IEEE (2003), IEEE Standard for Interconnecting Distributed Resources with Electric Power Systems, Std. 1547-2003. IEEE (2004), Impact of Distributed Resources on Distribution Relay Protection, Special report of the IEEE, Power System Relay Committee Working Group D3, August 2004. Jenkins, N.; Allan, R.; Crossley, P.; Kirschen, D. & Strbac, G. (2000), Embedded Generation, IEE Power and Energy Series 13, 2000. Jiang, S.; Annakkage, U.D. & Gole, A. M. (2005) , A platform for validation of FACTS models, IEEE Trans. Power Delivery, Vol. 20, No.4, October 2005. Kane, P.O. & Fox, B. (1997), Loss of Mains Detection for Embedded Generation by System Impedance Monitoring Developments in Power System Protection, IEE Conference Publication No. 434, 25-27th March 1997, pp: 95-98. Lopes, J.A.P. (2002), Integration of dispersed generation on distribution networks impact studies, IEEE Power Engineering Society Winter Meeting, Vol. 1, pages 323328, 2002. Mozina, C.J. (2001), Interconnection protection of IPP generators at commercial/industrial facilities, IEEE Transactions on Industry Applications, pp 681-688, 2001. Peters, R.R.; Muthumuni, D.; Bartel, T.; Salehfar, H. & Mann, M. (2006), Dynamic model development of a fixed speed stall control wind turbine at start-up, IEEE Power Engineering Society General Meeting, pp: June 2006, pp: 1-7. Redfern, M. A.; Usta, O. & Fielding, G. (1993), Protection Against Loss Of Utility Grid Supply For A Dispersed Storage And Generation Unit, IEEE Transactions on Power Delivery, Vol. 8, No. 3, July 1993, pp: 948-954. Ropp, M. E. ; Begovic, M. & Rohatgi, A. (1999), Analysis and Performance Assessment of the Active Frequency Drift Method of Islanding Prevention, IEEE Transactions on Energy Conversion, Vol. 14, No. 3, September 1999, pp: 810–816. Rudion, K.; Orths, A.; Styczynski, Z. & Strunz K.,(2005), Design of benchmark of medium voltage distribution network for investigation of DG integration, Proc. Conference on the Innovations, Commercial Applications and Regulatory Frameworks of Distributed Generation. November 30, December 1-2, 2005, Boston, USA. Uchida, A.; Watanabe,S. & Iwamoto, S. (2006), A Voltage Control Strategy for Distribution Networks with Dispersed Generations, Proceeding of IEEE – PES General Meeting, pp. 1-6, Tampa, Florida, USA, June 2007. Grid Integration of Renewable Energy Systems 131Renewable Energy132 Hardware in the loop simulation of renewable distributed generation systems 133Hardware in the loop simulation of renewable distributed generation systemsMarco MauriX Hardware in the loop simulation of renewable distributed generation systems Marco Mauri Department of Mechanics, Politecnico di Milano Italy 1. Introduction The strong growth of Distributed Generation (DG) systems installed on Low Voltage (LV) electrical power networks (with a typical power range of 1 to 400 kW) has emphasized the need to perform preliminary tests before they are put into service to check their efficiency and validate Power Quality (PQ) functions. Typically, these tests are made on DG system prototypes and entail high experimental cost and effort. In addition, it is often very hard to perform significant tests in a short time because DG systems use renewable energy sources, like the wind or sun that are not always available. A hybrid approach, using Hardware in the Loop (HIL) simulation, has turned out to be an effective tool for the evaluation of these systems and various methods have been proposed in the literature, where Hardware in the Loop (HIL) simulation refers to a system in which parts of a pure simulation have been replaced with actual physical components. Hardware in the loop simulations (HILS) are not new, but have been used extensively for controller assessment for a long time. The aerospace industry has been using this technique ever since software became a safety critical aspect of flight control systems. More recently, however, a combination of several factors has led to a sharp increase in its use, principally the reduced cost and greater availability of HILS products and the intense pressure to reduce development time and costs. HIL simulation is now used more and more to develop new components and actuators in many fields. Vehicle component evaluation, drive control assessment, power electronics and electric grids, servo control and robotics, railway traction systems for trains and subways and educational applications can all be cited. More recently, HIL simulations have also been used to test distributed generation systems, e.g., to validate control algorithms or Maximum Power Point Tracking (MPPT) strategies. Generally, these HIL techniques are limited to control applications, i.e., there is only signal coupling between the hardware and the virtual system. This approach, even if it is very useful during the control design stage, presents some deficiencies when it is used to develop electrical drives for renewable low power energy generators. In these systems, it is very important to evaluate the interactions between the electrical drives and the mechanical components in order to avoid dangerous working conditions and, moreover, to estimate precisely the system’s energy efficiency. In particular, this parameter is 8Renewable Energy132 Hardware in the loop simulation of renewable distributed generation systems 133Hardware in the loop simulation of renewable distributed generation systemsMarco MauriX Hardware in the loop simulation of renewable distributed generation systems Marco Mauri Department of Mechanics, Politecnico di Milano Italy 1. Introduction The strong growth of Distributed Generation (DG) systems installed on Low Voltage (LV) electrical power networks (with a typical power range of 1 to 400 kW) has emphasized the need to perform preliminary tests before they are put into service to check their efficiency and validate Power Quality (PQ) functions. Typically, these tests are made on DG system prototypes and entail high experimental cost and effort. In addition, it is often very hard to perform significant tests in a short time because DG systems use renewable energy sources, like the wind or sun that are not always available. A hybrid approach, using Hardware in the Loop (HIL) simulation, has turned out to be an effective tool for the evaluation of these systems and various methods have been proposed in the literature, where Hardware in the Loop (HIL) simulation refers to a system in which parts of a pure simulation have been replaced with actual physical components. Hardware in the loop simulations (HILS) are not new, but have been used extensively for controller assessment for a long time. The aerospace industry has been using this technique ever since software became a safety critical aspect of flight control systems. More recently, however, a combination of several factors has led to a sharp increase in its use, principally the reduced cost and greater availability of HILS products and the intense pressure to reduce development time and costs. HIL simulation is now used more and more to develop new components and actuators in many fields. Vehicle component evaluation, drive control assessment, power electronics and electric grids, servo control and robotics, railway traction systems for trains and subways and educational applications can all be cited. More recently, HIL simulations have also been used to test distributed generation systems, e.g., to validate control algorithms or Maximum Power Point Tracking (MPPT) strategies. Generally, these HIL techniques are limited to control applications, i.e., there is only signal coupling between the hardware and the virtual system. This approach, even if it is very useful during the control design stage, presents some deficiencies when it is used to develop electrical drives for renewable low power energy generators. In these systems, it is very important to evaluate the interactions between the electrical drives and the mechanical components in order to avoid dangerous working conditions and, moreover, to estimate precisely the system’s energy efficiency. In particular, this parameter is 8Renewable Energy134 an important variable in a renewable energy plant because these systems are quite expensive and a small increase in global efficiency can drastically reduce the payback time. For this reason, different kinds of HIL simulations can be used to meet these requirements. 2. HIL simulation concepts for Distributed Generation systems DG systems can present different topologies, but a wide number of them require that a static power converter be efficiently connected to an LV network or that active power with a stable voltage and frequency is provided. Indeed, the power converter interface can perform different functions:  Compensation of reactive power (with or without storage elements).  Compensation of negative sequence components and components with different frequencies for the current.  Compensation of voltage sags, micro interruptions and interruptions (with storage elements). DG sources can rely on both renewable and non-renewable sources, such as wind turbines, photovoltaic generators, small hydro, fuel cells and gas powered combined heat and power (CHP) stations. Figure 1 shows a simplified schematic of a complete DG system. AC DCField circuit SG GE Sensitive load Insensitive load Thermal load Dissipative loadDC AC Y HMI DC DCDC DCPV system Public network Braking chopper Å UDCGE: Gas Engine SG: Synchronous GeneratorDG SOURCE Fig. 1. The simplified architecture of a generic Distributed Generation (DG) system. The Electrical Drive constitutes the main subsystem of a renewable DG system and the major development effort is often concentrated on this part. An electrical drive can be decomposed into several subsystems: the control software, the power electronics set, the electrical machine and the mechanical load to move (Fig. 2). Typically, a dedicated control board implements the control software and supplies the switching commands to the power electronics converter. Measurements of all the power parts are inputs for the controller board. Power Electronics Electrical Machine Mechanical Load/Gen. Process Control Measurement Switching signals Fig. 2. The Electrical Drive structure. In a traditional computer simulation, all of these parts are simulated in the same simulation software environment and, in order to reduce the computation time, simple models and other simplifications are considered (e.g., the sampling period of the control is often neglected). This approach is not always accurate enough to allow a real-time implementation of the control without other intermediate steps. HIL simulation could be a very useful intermediary step: one of the simulated parts can be replaced by its hardware component in order to take into account the real constraints in the simulation loop (Fig. 3). Software Simulation HIL tests Hardware implementation Fig. 3. Electrical Drive design procedure. Different kinds of HIL simulations can be considered:  Signal level HIL simulation, where only the control board is tested and the other parts are simulated in real-time. This kind of HIL simulation has very often been employed in aerospace and automotive applications for the assessment of controller boards (Fig. 4).  Power level HIL simulation, where the control board and the power electronics converter are evaluated and the electrical machine and the mechanical load are simulated (Fig. 5).  Mechanical level HIL simulation. In this case, the whole drive (control, power electronics and electrical machine) is tested and the mechanical part is simulated (Fig. 6). Each of these simulations can be used to test a DG electrical drive, although obviously highlighting different aspects and problems. 2.1 Signal Level HIL simulation In a Signal Level Hardware in the Loop Simulation, only the controller board (which contains the process control) is tested (Fig. 4). The other parts (power electronics, machine and mechanical load) are simulated in real-time. The simulation system must manage the inputs and outputs of the controller board under test. A second controller board is thus used to simulate in real-time the power parts of the system. Specific signal conditioning is required to impose the same inputs and outputs as imposed by the power parts. This method can be called “signal level HIL simulation” because only signals are used at the Hardware in the loop simulation of renewable distributed generation systems 135Renewable Energy134 an important variable in a renewable energy plant because these systems are quite expensive and a small increase in global efficiency can drastically reduce the payback time. For this reason, different kinds of HIL simulations can be used to meet these requirements. 2. HIL simulation concepts for Distributed Generation systems DG systems can present different topologies, but a wide number of them require that a static power converter be efficiently connected to an LV network or that active power with a stable voltage and frequency is provided. Indeed, the power converter interface can perform different functions:  Compensation of reactive power (with or without storage elements).  Compensation of negative sequence components and components with different frequencies for the current.  Compensation of voltage sags, micro interruptions and interruptions (with storage elements). DG sources can rely on both renewable and non-renewable sources, such as wind turbines, photovoltaic generators, small hydro, fuel cells and gas powered combined heat and power (CHP) stations. Figure 1 shows a simplified schematic of a complete DG system. AC DCField circuit SG GE Sensitive load Insensitive load Thermal load Dissipative loadDC AC Y HMI DC DCDC DCPV system Public network Braking chopper Å UDCGE: Gas Engine SG: Synchronous GeneratorDG SOURCE Fig. 1. The simplified architecture of a generic Distributed Generation (DG) system. The Electrical Drive constitutes the main subsystem of a renewable DG system and the major development effort is often concentrated on this part. An electrical drive can be decomposed into several subsystems: the control software, the power electronics set, the electrical machine and the mechanical load to move (Fig. 2). Typically, a dedicated control board implements the control software and supplies the switching commands to the power electronics converter. Measurements of all the power parts are inputs for the controller board. Power Electronics Electrical Machine Mechanical Load/Gen. Process Control Measurement Switching signals Fig. 2. The Electrical Drive structure. In a traditional computer simulation, all of these parts are simulated in the same simulation software environment and, in order to reduce the computation time, simple models and other simplifications are considered (e.g., the sampling period of the control is often neglected). This approach is not always accurate enough to allow a real-time implementation of the control without other intermediate steps. HIL simulation could be a very useful intermediary step: one of the simulated parts can be replaced by its hardware component in order to take into account the real constraints in the simulation loop (Fig. 3). Software Simulation HIL tests Hardware implementation Fig. 3. Electrical Drive design procedure. Different kinds of HIL simulations can be considered:  Signal level HIL simulation, where only the control board is tested and the other parts are simulated in real-time. This kind of HIL simulation has very often been employed in aerospace and automotive applications for the assessment of controller boards (Fig. 4).  Power level HIL simulation, where the control board and the power electronics converter are evaluated and the electrical machine and the mechanical load are simulated (Fig. 5).  Mechanical level HIL simulation. In this case, the whole drive (control, power electronics and electrical machine) is tested and the mechanical part is simulated (Fig. 6). Each of these simulations can be used to test a DG electrical drive, although obviously highlighting different aspects and problems. 2.1 Signal Level HIL simulation In a Signal Level Hardware in the Loop Simulation, only the controller board (which contains the process control) is tested (Fig. 4). The other parts (power electronics, machine and mechanical load) are simulated in real-time. The simulation system must manage the inputs and outputs of the controller board under test. A second controller board is thus used to simulate in real-time the power parts of the system. Specific signal conditioning is required to impose the same inputs and outputs as imposed by the power parts. This method can be called “signal level HIL simulation” because only signals are used at the Hardware in the loop simulation of renewable distributed generation systems 135Renewable Energy136 interface between the system under test and the simulation environment. This kind of HIL has very often been employed in aerospace and automotive applications for the assessment of controller boards. System under test control Simulation environment Real-time simulation Fig. 4. Signal Level HIL simulation. 2.2 Power Level HIL simulation In a Power Level HIL simulation, the actual controller board and the power electronics converter are evaluated. The other parts (electrical machine and mechanical load) are simulated. The simulation system must impose inputs and outputs for the power electronics and the controller board under test. The simulation environment is generally composed of a second power electronics set (electric load) and a second controller board (real-time simulation) (Fig. 5). System under test control Simulation environment Real-time simulation Electronic load Power converter Fig. 5. Power Level HIL simulation. 2.3 Mechanical Level HIL simulation In this case, the whole drive (control, power electronics and electric machine) is tested and the mechanical part is simulated. The simulation system must impose mechanical inputs and outputs on the electrical machine under test. Moreover, measurements on the mechanical part have to be sent to the controller board under test. Another electrical machine (load machine) is often used as a controlled mechanical load. It is supplied by a second power electronics set (load supply). A second controller board (real-time simulation) is required to control the load machine and to send fictitious mechanical "measurements" to the controller board under test (Fig. 6). System under test control Simulation environment Real-time simulation Electrical Machine Power Converter Electrical Machine Power Converter Fig. 6. Mechanical Level HIL simulation. 3. Real Time modeling concepts The implementation of real time models of controlled devices is a well-known problem but becomes a more difficult challenge when the time constraints are very narrow. The simulation models should satisfy different requirements:  precision and reliability to represent physical system behavior;  a simple mathematical structure to allow for fast numerical integration and easy software implementation;  a standard structure for the model equation in order to obtain a more flexible and easily configurable simulation platform. An Object Oriented approach is typically followed to meet these characteristics during the implementation of emulation models, making it possible to improve the emulator flexibility. Each part of the complete system has been considered as a single object with inputs, outputs and internal states. In this way it is very simple to reconfigure the simulation system without changing the hardware and software structure. 3.1 Signal Level Modeling DC bridge i1(n) i2(n) + - v(n-1) Inverter v(n -1) + - i1(n-1) i2(n-1) v(n) Fig. 7. Example of an Object-Oriented integration procedure for the DC Stage. In signal level real-time modeling, the smallest bound is the time step necessary to implement the high dynamic behavior of the power electronics stage. This time step is very small (typically 10-100 s), so it is necessary to make a great effort to develop suitable models. Hardware in the loop simulation of renewable distributed generation systems 137Renewable Energy136 interface between the system under test and the simulation environment. This kind of HIL has very often been employed in aerospace and automotive applications for the assessment of controller boards. System under test control Simulation environment Real-time simulation Fig. 4. Signal Level HIL simulation. 2.2 Power Level HIL simulation In a Power Level HIL simulation, the actual controller board and the power electronics converter are evaluated. The other parts (electrical machine and mechanical load) are simulated. The simulation system must impose inputs and outputs for the power electronics and the controller board under test. The simulation environment is generally composed of a second power electronics set (electric load) and a second controller board (real-time simulation) (Fig. 5). System under test control Simulation environment Real-time simulation Electronic load Power converter Fig. 5. Power Level HIL simulation. 2.3 Mechanical Level HIL simulation In this case, the whole drive (control, power electronics and electric machine) is tested and the mechanical part is simulated. The simulation system must impose mechanical inputs and outputs on the electrical machine under test. Moreover, measurements on the mechanical part have to be sent to the controller board under test. Another electrical machine (load machine) is often used as a controlled mechanical load. It is supplied by a second power electronics set (load supply). A second controller board (real-time simulation) is required to control the load machine and to send fictitious mechanical "measurements" to the controller board under test (Fig. 6). System under test control Simulation environment Real-time simulation Electrical Machine Power Converter Electrical Machine Power Converter Fig. 6. Mechanical Level HIL simulation. 3. Real Time modeling concepts The implementation of real time models of controlled devices is a well-known problem but becomes a more difficult challenge when the time constraints are very narrow. The simulation models should satisfy different requirements:  precision and reliability to represent physical system behavior;  a simple mathematical structure to allow for fast numerical integration and easy software implementation;  a standard structure for the model equation in order to obtain a more flexible and easily configurable simulation platform. An Object Oriented approach is typically followed to meet these characteristics during the implementation of emulation models, making it possible to improve the emulator flexibility. Each part of the complete system has been considered as a single object with inputs, outputs and internal states. In this way it is very simple to reconfigure the simulation system without changing the hardware and software structure. 3.1 Signal Level Modeling DC bridge i1(n) i2(n) + - v(n-1) Inverter v(n -1) + - i1(n-1) i2(n-1) v(n) Fig. 7. Example of an Object-Oriented integration procedure for the DC Stage. In signal level real-time modeling, the smallest bound is the time step necessary to implement the high dynamic behavior of the power electronics stage. This time step is very small (typically 10-100 s), so it is necessary to make a great effort to develop suitable models. Hardware in the loop simulation of renewable distributed generation systems 137Renewable Energy138 However, a suitable choice for input/output object signals makes it possible to extend the Object Oriented approach during the simulation phase in order to improve the simulation speed because each object can be integrated separately and in a parallel way. It is obvious that this approximation is acceptable only if input/output object signals change slowly. Signals that satisfy this requirement are the state variables of each subsystem. Figure 7 shows the proposed procedure for the connection between two power converters of a DG system (Fig. 1). The presence of a capacitor on the DC bus makes it possible to decouple the integration of the two power converter models. Another big problem that is typical of this HIL simulation is the high sampling speed necessary in order to acquire control board switching commands. In order to decouple the acquisition time of the switching state and the integration, the average values of the switching functions are often used instead of the actual values. Different averaging methods are presented in the literature, according to the simulation requirements, especially in the emulation of system faults. An interesting averaging law is presented in (Bucca et al. 2006), where a two level inverter has been modeled using a dedicated acquisition board with two counters. In fact, the output voltage vector of the two-level inverter depends on the DC bus voltage value and on the gate signal sequence only (Fig. 8). The counter increment is a gate signal function (Table 1) and the mean values of the d and q components of the output voltage space vector could be obtained using Equation (1). D Q 100110010011001 101111000 Fig. 8. Output voltage vectors for a two level inverter. Gate signals swd i swq increment 000 0 0 001 1 2  3 2 010 21 23 011 1 0 100 1 0 101 21 23 110 21 23 111 0 0 Table 1. Counter increments. vsd 23VDC1Nswd (n)n1Nvsq 22VDC1Nswq (n)n1N (1) where:  N is the number of gate signal samples during an integration step;  swd(n), swq(n) are the n-th values of the counters according to Table 1;  VDC is the value of the DC bus voltage. Because this voltage is typically used as an input of electrical machine models that are written on a dq axis, this averaging method reduces the number of calculations, with a remarkable improvement in the simulation speed. 3.2 Power Level Modeling Power level HIL simulations have problems similar to Signal level HIL, because the time boundaries depend on the same electrical components. A big problem in a Power level HIL simulation is that it is necessary to decouple the dynamics of an electronic load from the power electronic converter. Typically, this simulation has been used in DG HIL simulations to test the PV field controller in order to also evaluate the power converter efficiency. A typical schematic of an HIL power level test bench is shown in Fig. 9. Control board Converter under test RLC filter Controlled inverter Fig. 9. Schematic of Power level HIL test bench. A three-phase RLC filter connects the hardware under test with the controlled converter. The filter decouples the operation of the converter from the hardware under test and reduces the ripple introduced by the switching operations of the controlled inverter. In this configuration, the currents drained in each branch of the filter can be regulated by controlling the output voltages of the inverter in order to track the references computed by the real-time simulation. Hardware in the loop simulation of renewable distributed generation systems 139Renewable Energy138 However, a suitable choice for input/output object signals makes it possible to extend the Object Oriented approach during the simulation phase in order to improve the simulation speed because each object can be integrated separately and in a parallel way. It is obvious that this approximation is acceptable only if input/output object signals change slowly. Signals that satisfy this requirement are the state variables of each subsystem. Figure 7 shows the proposed procedure for the connection between two power converters of a DG system (Fig. 1). The presence of a capacitor on the DC bus makes it possible to decouple the integration of the two power converter models. Another big problem that is typical of this HIL simulation is the high sampling speed necessary in order to acquire control board switching commands. In order to decouple the acquisition time of the switching state and the integration, the average values of the switching functions are often used instead of the actual values. Different averaging methods are presented in the literature, according to the simulation requirements, especially in the emulation of system faults. An interesting averaging law is presented in (Bucca et al. 2006), where a two level inverter has been modeled using a dedicated acquisition board with two counters. In fact, the output voltage vector of the two-level inverter depends on the DC bus voltage value and on the gate signal sequence only (Fig. 8). The counter increment is a gate signal function (Table 1) and the mean values of the d and q components of the output voltage space vector could be obtained using Equation (1). D Q 100110010011001 101111000 Fig. 8. Output voltage vectors for a two level inverter. Gate signals swd i swq increment 000 0 0 001 1 2  3 2 010 21 23 011 1 0 100 1 0 101 21 23 110 21 23 111 0 0 Table 1. Counter increments. vsd 23VDC1Nswd (n)n1Nvsq 22VDC1Nswq (n)n1N (1) where:  N is the number of gate signal samples during an integration step;  swd(n), swq(n) are the n-th values of the counters according to Table 1;  VDC is the value of the DC bus voltage. Because this voltage is typically used as an input of electrical machine models that are written on a dq axis, this averaging method reduces the number of calculations, with a remarkable improvement in the simulation speed. 3.2 Power Level Modeling Power level HIL simulations have problems similar to Signal level HIL, because the time boundaries depend on the same electrical components. A big problem in a Power level HIL simulation is that it is necessary to decouple the dynamics of an electronic load from the power electronic converter. Typically, this simulation has been used in DG HIL simulations to test the PV field controller in order to also evaluate the power converter efficiency. A typical schematic of an HIL power level test bench is shown in Fig. 9. Control board Converter under test RLC filter Controlled inverter Fig. 9. Schematic of Power level HIL test bench. A three-phase RLC filter connects the hardware under test with the controlled converter. The filter decouples the operation of the converter from the hardware under test and reduces the ripple introduced by the switching operations of the controlled inverter. In this configuration, the currents drained in each branch of the filter can be regulated by controlling the output voltages of the inverter in order to track the references computed by the real-time simulation. Hardware in the loop simulation of renewable distributed generation systems 139Renewable Energy140 A critical step in the realization of this test bench is the measurement of the voltages generated by the converter under test, because these voltages represent the inputs for the real-time simulation. A way to overcome this problem is to estimate these voltages by acquiring the switch signal in a way similar to the signal level HIL test bench, but this structure depends on the control electronics of the converter under test. In fact, the platform interface has to be specifically customized based on the characteristics of the signals that drive the switch gates. In addition, these signals may not be available in industrial converters. The easiest solution if the gate signals cannot be used is to measure the average voltage of the converter output directly. The voltages are scaled and reduced to a signal level by means of suitable sensors (e.g., hall sensors) that provide galvanic isolation, and then they are averaged over a simulation period. The main advantage of this technique is that the voltage used in the real-time simulation is reconstructed from the power signal instead of the gate signal in order to include switching non-linearity and improve the HIL test bench standardization. 3.3 Mechanical Level Modeling In a mechanical level HIL simulation, the presence of a real electrical drive removes the very small time step boundaries of signal level and power level HIL simulations and introduces new boundaries that depend on mechanical dynamics (2). Tm  Tr  Jddt, (2) where Tm is the electromagnetic torque, Tr is the load torque and J is the equivalent inertia. Equation (2) shows that the electrical drive used to emulate the mechanical load should supply not only the load torque but also the inertial torque, which could present high values and dynamics. To satisfy this requirement, it is necessary to use a high performance drive, which is not very simple to implement and control. To avoid this problem, a different approach can be used, as proposed in (). In particular, the DG power source presents a low dynamic speed variation due to a high damping factor and inertia, so it is possible to modify the mechanical emulator control of the electrical drive from torque to speed control. In this way, the mechanical model could be reduced to a quasi-steady-state model by considering the DG power source speed as a constant during the integration step. This is possible because the integration model step is very small compared with the mechanical time constants, due to the previous considerations. Using this approach, the acceleration torque can be calculated by subtracting the load power (Pl) from the available mechanical source power (Ps) and dividing the result by the actual speed. The load power is measured by the electrical drive to take into account all the mechanical losses in the power train. The new speed reference (ωref) for the electrical drive that emulates the DG power source is calculated by dividing the acceleration torque by the system equivalent inertia, as is indicated in (3) and (4) and represented in Fig. 10. Ta Ps  Pl act (3) d(m)dt1JTacc Bact (4) m TaJBactdt0t  ref  act  m (5) Mechanical model Speed ( ) 1Electrical drive torque - Acceleration torque 1JSpeed reference + Fig. 10. Calculation of the new speed reference. The proposed approach simplifies the modeling of high inertia loads, making it possible to test control strategies under dynamic conditions, while avoiding the measurement of load acceleration. Moreover, it is possible to use a less sophisticated electrical drive, control board and emulation board, thus reducing the test bench cost. A different approach was proposed in (Kuperman and Rabinovici, 2005), where the simulation of load variation was introduced by adding a virtual load signal to the output of the electrical motor controller. Even though this approach is very interesting, it is very difficult to implement it, because it is necessary to modify the controller software. 4. Experimental results In the following sections, some experimental results using different HIL configurations are presented. Motor Rated power at 50 Hz 370 W Rated torque 1.3 Nm Rated speed 2820 rpm Rated current 1.7 A Starting current 4.56 A Power factor 0.83 Starting Torque 3.9 Nm Rotor inertia 3.5  10-4 kgm2 Nr. of pole pairs 1 Statoric phase resistance 24.6  Rotoric phase resistance 16.1  Magnetizing inductance 1.46 H Electrical Time constant 1.62  10-3 s Hardware in the loop simulation of renewable distributed generation systems 141Renewable Energy140 A critical step in the realization of this test bench is the measurement of the voltages generated by the converter under test, because these voltages represent the inputs for the real-time simulation. A way to overcome this problem is to estimate these voltages by acquiring the switch signal in a way similar to the signal level HIL test bench, but this structure depends on the control electronics of the converter under test. In fact, the platform interface has to be specifically customized based on the characteristics of the signals that drive the switch gates. In addition, these signals may not be available in industrial converters. The easiest solution if the gate signals cannot be used is to measure the average voltage of the converter output directly. The voltages are scaled and reduced to a signal level by means of suitable sensors (e.g., hall sensors) that provide galvanic isolation, and then they are averaged over a simulation period. The main advantage of this technique is that the voltage used in the real-time simulation is reconstructed from the power signal instead of the gate signal in order to include switching non-linearity and improve the HIL test bench standardization. 3.3 Mechanical Level Modeling In a mechanical level HIL simulation, the presence of a real electrical drive removes the very small time step boundaries of signal level and power level HIL simulations and introduces new boundaries that depend on mechanical dynamics (2). Tm  Tr  Jddt, (2) where Tm is the electromagnetic torque, Tr is the load torque and J is the equivalent inertia. Equation (2) shows that the electrical drive used to emulate the mechanical load should supply not only the load torque but also the inertial torque, which could present high values and dynamics. To satisfy this requirement, it is necessary to use a high performance drive, which is not very simple to implement and control. To avoid this problem, a different approach can be used, as proposed in (). In particular, the DG power source presents a low dynamic speed variation due to a high damping factor and inertia, so it is possible to modify the mechanical emulator control of the electrical drive from torque to speed control. In this way, the mechanical model could be reduced to a quasi-steady-state model by considering the DG power source speed as a constant during the integration step. This is possible because the integration model step is very small compared with the mechanical time constants, due to the previous considerations. Using this approach, the acceleration torque can be calculated by subtracting the load power (Pl) from the available mechanical source power (Ps) and dividing the result by the actual speed. The load power is measured by the electrical drive to take into account all the mechanical losses in the power train. The new speed reference (ωref) for the electrical drive that emulates the DG power source is calculated by dividing the acceleration torque by the system equivalent inertia, as is indicated in (3) and (4) and represented in Fig. 10. Ta Ps  Pl act (3) d(m)dt1JTacc Bact (4) m TaJBactdt0t  ref  act  m (5) Mechanical model Speed ( ) 1Electrical drive torque - Acceleration torque 1JSpeed reference + Fig. 10. Calculation of the new speed reference. The proposed approach simplifies the modeling of high inertia loads, making it possible to test control strategies under dynamic conditions, while avoiding the measurement of load acceleration. Moreover, it is possible to use a less sophisticated electrical drive, control board and emulation board, thus reducing the test bench cost. A different approach was proposed in (Kuperman and Rabinovici, 2005), where the simulation of load variation was introduced by adding a virtual load signal to the output of the electrical motor controller. Even though this approach is very interesting, it is very difficult to implement it, because it is necessary to modify the controller software. 4. Experimental results In the following sections, some experimental results using different HIL configurations are presented. Motor Rated power at 50 Hz 370 W Rated torque 1.3 Nm Rated speed 2820 rpm Rated current 1.7 A Starting current 4.56 A Power factor 0.83 Starting Torque 3.9 Nm Rotor inertia 3.5  10-4 kgm2 Nr. of pole pairs 1 Statoric phase resistance 24.6  Rotoric phase resistance 16.1  Magnetizing inductance 1.46 H Electrical Time constant 1.62  10-3 s Hardware in the loop simulation of renewable distributed generation systems 141Renewable Energy142 Converter Input voltage single-phase 180-240 V Output power 750W, 150% overl.1 min. Output voltage 0-230 V Power switches IGBT Table 2. Emulated Converter and Motor characteristics. 4.1 Signal Level HIL simulation An example of a real-time emulator hardware structure for signal level HIL simulations is presented in Fig. 11. The HIL test bench emulates an asynchronous drive with the characteristics indicated in Table 2. The electrical drive used for the test was a Technosoft induction motor drive with a control board based on TMS320F240 DSP. Dspace Boards Interface Board Control Board speed and current PWM signal Encoder and current signals PWM signal Digital control signal s Fig. 11. Signal level HIL test bench architecture. A dSpace system (ds1006 board based) was used to integrate the electrical drive models, acquire the PWM signal and generate speed and current signals. A dedicated interface board was used to adapt the analogue signal levels and generate the encoder signals. Speed V/Hz 0 1 Preset Preset Enc 1 Enc 2 0-100kHz 0-10V Signals conditioning Fig. 12. Encoder emulation. The encoder signal generation was done using a Voltage/Frequency converter and two digital counters (Fig. 12). The basic idea was based on the observation that if the counter chips have been initialized to 0 and 1, respectively, the output 2nd bit signals are square signals with a π/2 phase shift. Fig. 13 shows the phase current and speed signals acquired by the control board for different speed transients. The induction motor was controlled using close loop field oriented control. a) Phase current b) Speed c) Phase current d) Speed Fig. 13. Comparison between real motor acquisition and emulated system acquisition of phase current and speed for different speed transients. It is interesting to note that the real transients and emulated transients are very similar. There was a little difference in the first speed acquisition (Fig. 13b), but it was practically limited to the less significant bit of the digital acquisition. This difference was principally caused by the absence of any additional conditioning stage between the encoder emulation board and the control board and was more significant at low speed (comparing Fig. 13b-d). 4.2 Power Level HIL simulation A power level HIL simulation of a PV field was performed. A typical configuration for a photovoltaic (PV) generator is shown in Fig. 14. Hardware in the loop simulation of renewable distributed generation systems 143Signals conditioning Renewable Energy142 Converter Input voltage single-phase 180-240 V Output power 750W, 150% overl.1 min. Output voltage 0-230 V Power switches IGBT Table 2. Emulated Converter and Motor characteristics. 4.1 Signal Level HIL simulation An example of a real-time emulator hardware structure for signal level HIL simulations is presented in Fig. 11. The HIL test bench emulates an asynchronous drive with the characteristics indicated in Table 2. The electrical drive used for the test was a Technosoft induction motor drive with a control board based on TMS320F240 DSP. Dspace Boards Interface Board Control Board speed and current PWM signal Encoder and current signals PWM signal Digital control signal s Fig. 11. Signal level HIL test bench architecture. A dSpace system (ds1006 board based) was used to integrate the electrical drive models, acquire the PWM signal and generate speed and current signals. A dedicated interface board was used to adapt the analogue signal levels and generate the encoder signals. Speed V/Hz 0 1 Preset Preset Enc 1 Enc 2 0-100kHz 0-10V Signals conditioning Fig. 12. Encoder emulation. The encoder signal generation was done using a Voltage/Frequency converter and two digital counters (Fig. 12). The basic idea was based on the observation that if the counter chips have been initialized to 0 and 1, respectively, the output 2nd bit signals are square signals with a π/2 phase shift. Fig. 13 shows the phase current and speed signals acquired by the control board for different speed transients. The induction motor was controlled using close loop field oriented control. a) Phase current b) Speed c) Phase current d) Speed Fig. 13. Comparison between real motor acquisition and emulated system acquisition of phase current and speed for different speed transients. It is interesting to note that the real transients and emulated transients are very similar. There was a little difference in the first speed acquisition (Fig. 13b), but it was practically limited to the less significant bit of the digital acquisition. This difference was principally caused by the absence of any additional conditioning stage between the encoder emulation board and the control board and was more significant at low speed (comparing Fig. 13b-d). 4.2 Power Level HIL simulation A power level HIL simulation of a PV field was performed. A typical configuration for a photovoltaic (PV) generator is shown in Fig. 14. Hardware in the loop simulation of renewable distributed generation systems 143Signals conditioning Renewable Energy144 Fig. 14. Configuration of emulated PV field. Normally a set of PV modules is series and parallel connected to form a PV multiple module or PV subfield, with suitable rated voltage and current. The PV subfield is then connected to a DC-DC converter. It is also possible to connect individual PV strings to their own DC-DC converters. The PV generator can work either in a grid connection mode or in stand alone operation without a storage system, thanks to a suitable control strategy. When a photovoltaic system works in the grid connection mode, a MPPT control technique is implemented in order to extract the maximum electrical energy from the PV panel. Using the HIL test bench, a DC/DC converter for a PV field with a Perturb and Observe MPPT algorithm was tested. The behavior of a SHELL SQ8 PV panel was emulated. Fig. 15 shows the behavior of a SHELL SQ8 PV panel for different irradiance levels and temperatures. 0 5 10 15 2005101520253035404550S=600W/m 2 S=300W/m 2 data5data60 5 10 15 2005101520253035404550T=330KT=360Kdata4data5data6 Fig. 15. SHELL Q8 PV V/I characteristics. The implemented P&O algorithm operates by periodically perturbing (i.e., incrementing or decrementing) the array terminal voltage and comparing the PV output power with that of the previous perturbation cycle. If the PV array operating voltage changes and the power increases (dP/dVPV>0), the control system moves the PV array operating point in that direction; otherwise the operating point is moved in the opposite direction. In the next perturbation cycle, the algorithm continues in the same way. A common problem in P&O algorithms is that the array terminal voltage is perturbed every MPPT cycle; therefore when the MPP is reached, the output power oscillates around the maximum, resulting in power loss in the PV system. This is especially true in constant or slowly-varying atmospheric conditions, but also occurs under rapidly changing atmospheric conditions (in this case the perturbation direction can be wrong). There are many different P&O methods available in the scientific literature. In this case, a modified P&O algorithm was adopted in which the perturbation amplitude of the PV operating point has a constantly variable magnitude. The magnitude of the perturbation is a function of the power variation necessary to reach the target. a) b) Fig. 16. a) 1 kW PV subfield I/V output characteristics during start up phase (I=20A => 5V; V=300V=>5V) b) I/V characteristics during start up phase of MPPT control. a) b) Fig. 17. 1 kW PV subfield I/V output characteristics when variable step P&O algorithm is used (I=20A => 5V; V=300V=>5V), a) during power discharging (from 1000W to 400 W) and b) during power changing (from 400 W to 1000 W). Fig. 16 presents the V/I characteristics during the PV subfield start up simulation. The PV subfield current starts from a null value with the maximum voltage value and reaches the MPP in 30 seconds, with constant current ripple amplitude. An instantaneous reduction in the PV subfield power from 1000 W to 400 W was simulated. The MPP changes from 74V-13A to 69V-5,5A. Fig. 17 represents the algorithm behavior during an instantaneous increase from 400 W to 1000 W. The MPP changes from 69V-5,5A to 74V-13A. Hardware in the loop simulation of renewable distributed generation systems 145Renewable Energy144 Fig. 14. Configuration of emulated PV field. Normally a set of PV modules is series and parallel connected to form a PV multiple module or PV subfield, with suitable rated voltage and current. The PV subfield is then connected to a DC-DC converter. It is also possible to connect individual PV strings to their own DC-DC converters. The PV generator can work either in a grid connection mode or in stand alone operation without a storage system, thanks to a suitable control strategy. When a photovoltaic system works in the grid connection mode, a MPPT control technique is implemented in order to extract the maximum electrical energy from the PV panel. Using the HIL test bench, a DC/DC converter for a PV field with a Perturb and Observe MPPT algorithm was tested. The behavior of a SHELL SQ8 PV panel was emulated. Fig. 15 shows the behavior of a SHELL SQ8 PV panel for different irradiance levels and temperatures. 0 5 10 15 2005101520253035404550S=600W/m 2 S=300W/m 2 data5data60 5 10 15 2005101520253035404550T=330KT=360Kdata4data5data6 Fig. 15. SHELL Q8 PV V/I characteristics. The implemented P&O algorithm operates by periodically perturbing (i.e., incrementing or decrementing) the array terminal voltage and comparing the PV output power with that of the previous perturbation cycle. If the PV array operating voltage changes and the power increases (dP/dVPV>0), the control system moves the PV array operating point in that direction; otherwise the operating point is moved in the opposite direction. In the next perturbation cycle, the algorithm continues in the same way. A common problem in P&O algorithms is that the array terminal voltage is perturbed every MPPT cycle; therefore when the MPP is reached, the output power oscillates around the maximum, resulting in power loss in the PV system. This is especially true in constant or slowly-varying atmospheric conditions, but also occurs under rapidly changing atmospheric conditions (in this case the perturbation direction can be wrong). There are many different P&O methods available in the scientific literature. In this case, a modified P&O algorithm was adopted in which the perturbation amplitude of the PV operating point has a constantly variable magnitude. The magnitude of the perturbation is a function of the power variation necessary to reach the target. a) b) Fig. 16. a) 1 kW PV subfield I/V output characteristics during start up phase (I=20A => 5V; V=300V=>5V) b) I/V characteristics during start up phase of MPPT control. a) b) Fig. 17. 1 kW PV subfield I/V output characteristics when variable step P&O algorithm is used (I=20A => 5V; V=300V=>5V), a) during power discharging (from 1000W to 400 W) and b) during power changing (from 400 W to 1000 W). Fig. 16 presents the V/I characteristics during the PV subfield start up simulation. The PV subfield current starts from a null value with the maximum voltage value and reaches the MPP in 30 seconds, with constant current ripple amplitude. An instantaneous reduction in the PV subfield power from 1000 W to 400 W was simulated. The MPP changes from 74V-13A to 69V-5,5A. Fig. 17 represents the algorithm behavior during an instantaneous increase from 400 W to 1000 W. The MPP changes from 69V-5,5A to 74V-13A. Hardware in the loop simulation of renewable distributed generation systems 145Renewable Energy146 4.3 Mechanical Level HIL simulation Two different DG systems were emulated using an HIL mechanical level test bench.  small CHP Generating Set  Small-scale hydropower generator 4.3.1 Small CHP generating set. Small power CHP (Combined Heat and Power) generating sets, based on internal combustion engines, are normally grid-connected and have a power range of 10 kW to 100 kW. These small CHP generating sets can be equipped with both diesel engines and gas fuelled engines. The advantage of using combined heat and power generation is that the global efficiency can be increased to 85 to 90%, which make them a very favorable solution for Distributed Generation systems. The most convenient arrangement is to use an asynchronous generator, even if it shows some drawbacks. In particular, it requires reactive power for the excitation of its magnetic circuits and, in the case of failure of the mains, it does not work in a stand alone mode to provide electrical power to the on-site ac load. Conventional CHP sets work at a fixed speed in order to generate electrical power at a fixed frequency, but they show low energy efficiency when they operate at partial load. An improvement can be obtained by using a variable speed gas engine to drive the electrical generator at a variable frequency. Operating the CHP set at a variable speed makes it possible to achieve the highest energy conversion efficiency at any partial load of more than 20%, but it is necessary to use an AC/AC frequency converter in order to connect the variable speed CHP set to the public grid. The behavior of a variable speed CHP set can be further enhanced by introducing an additional DC/DC static converter, in order to increase the amount of electrical power that can be transformed into heat (dissipating electrical power on a resistor). In this way, a control strategy with heating priority can meet a temporary high demand for heat, reducing the production of electricity. 4.3.2 Small-scale hydropower generator Hydroelectric generation has been well recognized as environmentally friendly and socially beneficial for many applications, but large-scale hydroelectric plant developments have become more and more difficult because of the shortage of undeveloped suitable sites and concerns about global environmental protection. However, there are many sites where it is possible to install a small-scale hydroelectric plant. These systems normally adopt expensive Kaplan turbines that allow optimal control of the water flow. In order to reduce plant costs in small-scale plants, it is preferable to use fixed-blade propeller water turbines. This choice can reduce the plant power efficiency because typically in traditional hydroelectric power plants, the speed of the generating unit remains constant to keep it synchronized with the grid, neglecting turbine discharge variation over time. Moreover, this situation becomes critical in stand-alone operation mode, where keeping the speed of the generation unit constant reduces the power plant efficiency drastically. In these cases, a variable speed operation mode and maximum power efficiency operation point tracking is welcome, but it presents numerous design problems regarding the definition of the hardware architecture and power management strategies. Such small-scale hydroelectric plants generally use a PM synchronous machine as an electrical generator. 4.3.3 The HIL test bench Each of these applications presents similar problems. In particular, it is very important to evaluate the mechanical performance and, moreover, it is necessary to evaluate the global efficiency (converter and motor) with a good approximation. This it is very difficult to do using an HIL environment where the power converter is also emulated: the calculation effort necessary to simulate the electrical drive characteristics and losses does not allow simple and low cost real-time implementation. Because of these considerations, mechanical emulation is the better choice. An example of mechanical test bench hardware architecture is shown in Fig. 18, which includes:  A motor drive, including an inverter controlled squirrel cage induction motor. This motor was used to emulate the static and dynamic mechanical behavior of the DG power source. An incremental optical encoder was used to feed back the speed signal for control and emulation purposes.  A DC motor was used to emulate the generator side of the DG system. The rated parameters of the induction motor and generator are listed in Table I. Induction Motor DC Motor Rated voltage 380 V Rated voltage 240 V Rated current 35.1 A Rated current 60 A Rated Power 18.5 kW Rated power 14 kW Power factor 0.88 Rated speed 1465 rpm Rated speed 1500 rpm Wound connection D N° of poles 4 N° of poles 4 Table 3. Induction Motor and DC Motor rated Parameters. Hardware in the loop simulation of renewable distributed generation systems 147Renewable Energy146 4.3 Mechanical Level HIL simulation Two different DG systems were emulated using an HIL mechanical level test bench.  small CHP Generating Set  Small-scale hydropower generator 4.3.1 Small CHP generating set. Small power CHP (Combined Heat and Power) generating sets, based on internal combustion engines, are normally grid-connected and have a power range of 10 kW to 100 kW. These small CHP generating sets can be equipped with both diesel engines and gas fuelled engines. The advantage of using combined heat and power generation is that the global efficiency can be increased to 85 to 90%, which make them a very favorable solution for Distributed Generation systems. The most convenient arrangement is to use an asynchronous generator, even if it shows some drawbacks. In particular, it requires reactive power for the excitation of its magnetic circuits and, in the case of failure of the mains, it does not work in a stand alone mode to provide electrical power to the on-site ac load. Conventional CHP sets work at a fixed speed in order to generate electrical power at a fixed frequency, but they show low energy efficiency when they operate at partial load. An improvement can be obtained by using a variable speed gas engine to drive the electrical generator at a variable frequency. Operating the CHP set at a variable speed makes it possible to achieve the highest energy conversion efficiency at any partial load of more than 20%, but it is necessary to use an AC/AC frequency converter in order to connect the variable speed CHP set to the public grid. The behavior of a variable speed CHP set can be further enhanced by introducing an additional DC/DC static converter, in order to increase the amount of electrical power that can be transformed into heat (dissipating electrical power on a resistor). In this way, a control strategy with heating priority can meet a temporary high demand for heat, reducing the production of electricity. 4.3.2 Small-scale hydropower generator Hydroelectric generation has been well recognized as environmentally friendly and socially beneficial for many applications, but large-scale hydroelectric plant developments have become more and more difficult because of the shortage of undeveloped suitable sites and concerns about global environmental protection. However, there are many sites where it is possible to install a small-scale hydroelectric plant. These systems normally adopt expensive Kaplan turbines that allow optimal control of the water flow. In order to reduce plant costs in small-scale plants, it is preferable to use fixed-blade propeller water turbines. This choice can reduce the plant power efficiency because typically in traditional hydroelectric power plants, the speed of the generating unit remains constant to keep it synchronized with the grid, neglecting turbine discharge variation over time. Moreover, this situation becomes critical in stand-alone operation mode, where keeping the speed of the generation unit constant reduces the power plant efficiency drastically. In these cases, a variable speed operation mode and maximum power efficiency operation point tracking is welcome, but it presents numerous design problems regarding the definition of the hardware architecture and power management strategies. Such small-scale hydroelectric plants generally use a PM synchronous machine as an electrical generator. 4.3.3 The HIL test bench Each of these applications presents similar problems. In particular, it is very important to evaluate the mechanical performance and, moreover, it is necessary to evaluate the global efficiency (converter and motor) with a good approximation. This it is very difficult to do using an HIL environment where the power converter is also emulated: the calculation effort necessary to simulate the electrical drive characteristics and losses does not allow simple and low cost real-time implementation. Because of these considerations, mechanical emulation is the better choice. An example of mechanical test bench hardware architecture is shown in Fig. 18, which includes:  A motor drive, including an inverter controlled squirrel cage induction motor. This motor was used to emulate the static and dynamic mechanical behavior of the DG power source. An incremental optical encoder was used to feed back the speed signal for control and emulation purposes.  A DC motor was used to emulate the generator side of the DG system. The rated parameters of the induction motor and generator are listed in Table I. Induction Motor DC Motor Rated voltage 380 V Rated voltage 240 V Rated current 35.1 A Rated current 60 A Rated Power 18.5 kW Rated power 14 kW Power factor 0.88 Rated speed 1465 rpm Rated speed 1500 rpm Wound connection D N° of poles 4 N° of poles 4 Table 3. Induction Motor and DC Motor rated Parameters. Hardware in the loop simulation of renewable distributed generation systems 147Renewable Energy148 Fig. 18. Example of mechanical HIL test bench  A desktop PC with a dSpace 1104 board that implements the DG power source model. A dedicated interface board was used to acquire measurement signals from the field and the motor drive and to transmit control reference signals from the dSpace board to the motor drive.  A passive load to dissipate the generated power. 4.3.4 Experimental results According to the previous discussion, when simulating the gas engine of a CHP system, it is necessary to evaluate the equivalent inertia and damping factor, along with the generated power vs. rotational speed curve. In a gas engine, this curve is proportional to the throttle valve opening. Fig. 19. Generated power vs. speed and throttle valve opening. Fig. 19 represents the generated power vs. rotational speed for different throttle valve openings (in percentage) for a 30 kW gas engine. In order to validate different control strategies, it was also necessary to emulate the fuel consumption under different load conditions. Fig. 20 shows the fuel flow in kg/h vs. speed and the throttle valve openings for the same motor. Fig. 20. Fuel flow vs. speed and throttle valve opening. J  4 kg/m2B  3 J /(rad /s) (6) The estimated inertia and damping factor of the motor are indicated in (6). The fixed blade turbine of a small-scale hydropower plant can be modeled in the same way as the gas engine. In this case, the turbine discharge (Q) and efficiency (η) vs. speed are obtained using the steady-state curves and these values are used in combination with the turbine head (H) to calculate the available mechanical power (6). Table 4 shows the parameters used. Hardware in the loop simulation of renewable distributed generation systems 149Renewable Energy148 Fig. 18. Example of mechanical HIL test bench  A desktop PC with a dSpace 1104 board that implements the DG power source model. A dedicated interface board was used to acquire measurement signals from the field and the motor drive and to transmit control reference signals from the dSpace board to the motor drive.  A passive load to dissipate the generated power. 4.3.4 Experimental results According to the previous discussion, when simulating the gas engine of a CHP system, it is necessary to evaluate the equivalent inertia and damping factor, along with the generated power vs. rotational speed curve. In a gas engine, this curve is proportional to the throttle valve opening. Fig. 19. Generated power vs. speed and throttle valve opening. Fig. 19 represents the generated power vs. rotational speed for different throttle valve openings (in percentage) for a 30 kW gas engine. In order to validate different control strategies, it was also necessary to emulate the fuel consumption under different load conditions. Fig. 20 shows the fuel flow in kg/h vs. speed and the throttle valve openings for the same motor. Fig. 20. Fuel flow vs. speed and throttle valve opening. J  4 kg/m2B  3 J /(rad /s) (6) The estimated inertia and damping factor of the motor are indicated in (6). The fixed blade turbine of a small-scale hydropower plant can be modeled in the same way as the gas engine. In this case, the turbine discharge (Q) and efficiency (η) vs. speed are obtained using the steady-state curves and these values are used in combination with the turbine head (H) to calculate the available mechanical power (6). Table 4 shows the parameters used. Hardware in the loop simulation of renewable distributed generation systems 149Renewable Energy150 Phyd  HQg  (7) Rated power 3.2 kW Rated discharge 0.1 m3/s Design head 5.2 m Rated Rotational Speed 180 rpm Table 4. Hydro turbine technical data. J  10 kg/m2B  50 J /(rad /s) (8) The efficiency of the system depends on the opening of the distributing valve and the speed as indicated in Fig. 21, and the turbine speed is considered to have a constant value during the integration step (in this case it was set to 100 ms). Fig. 21. Efficiency vs. turbine speed and distributing valve opening. Fig. 22 shows the generated power reference vs. time and the real generated power vs. time using the hydropower turbine model, emulating a grid connection working mode. Fig. 13 compares the estimated generated power using the HIL hydropower model and the real power dissipated on the resistive load of the test bench. The power reference was calculated based on the load variations, while the generated power should also take into account the plant losses. These losses produced the stationary difference between the generated power and power reference. Fig. 22. Power reference and generated power vs. time. Fig. 23. Estimated generated power (from HIL test bench) and dissipated power on the resistive load. 5. Conclusion Hardware in the loop simulations (HILS) are a powerful method to test DG system components, such as control system electronics, without connection to the real equipment under control. The use of this technique does, however, raise challenges, such as complex interfacing requirements and the execution of an accurate equipment simulation that must operate in real-time. The increasing adoption of HILS has been driven by a number of factors, including the ability to perform system testing before the complete system is available and to simulate failure conditions that would be dangerous or expensive to create in a real plant. Hardware in the loop simulation of renewable distributed generation systems 151Renewable Energy150 Phyd  HQg  (7) Rated power 3.2 kW Rated discharge 0.1 m3/s Design head 5.2 m Rated Rotational Speed 180 rpm Table 4. Hydro turbine technical data. J  10 kg/m2B  50 J /(rad /s) (8) The efficiency of the system depends on the opening of the distributing valve and the speed as indicated in Fig. 21, and the turbine speed is considered to have a constant value during the integration step (in this case it was set to 100 ms). Fig. 21. Efficiency vs. turbine speed and distributing valve opening. Fig. 22 shows the generated power reference vs. time and the real generated power vs. time using the hydropower turbine model, emulating a grid connection working mode. Fig. 13 compares the estimated generated power using the HIL hydropower model and the real power dissipated on the resistive load of the test bench. The power reference was calculated based on the load variations, while the generated power should also take into account the plant losses. These losses produced the stationary difference between the generated power and power reference. Fig. 22. Power reference and generated power vs. time. Fig. 23. Estimated generated power (from HIL test bench) and dissipated power on the resistive load. 5. Conclusion Hardware in the loop simulations (HILS) are a powerful method to test DG system components, such as control system electronics, without connection to the real equipment under control. The use of this technique does, however, raise challenges, such as complex interfacing requirements and the execution of an accurate equipment simulation that must operate in real-time. The increasing adoption of HILS has been driven by a number of factors, including the ability to perform system testing before the complete system is available and to simulate failure conditions that would be dangerous or expensive to create in a real plant. Hardware in the loop simulation of renewable distributed generation systems 151Renewable Energy152 HILS is an important tool to reduce development time and ensure the safety and reliability of DG components. The availability of cost effective tools to perform HILS should lead to an increase in the use of this technique as a standard method to test DG components. 6. References Bouscayrol (2008). Different types of Hardware-In-the-Loop simulation for electric drives. Proceedings of IEEE International Symposium on Industrial Electronics, 2008. ISIE 2008. pp. 2146 – 2151, ISBN 978-1-4244-1665-3, Cambridge, July 2008, IEEE Bin Lu; Xin Wu; Figueroa, H. & Monti, A. (2007). A Low-Cost Real-Time Hardware-in-the-Loop Testing Approach of Power Electronics Controls. Industrial Electronics, IEEE Transactions on, vol. 54 n. 2, pp. 919 – 931, ISSN 0278-0046 Bucca G.; Mapelli F. & Mauri M. (2006). A Real-time environment for industrial electrical drive control board testing. Proceedings of International Conference on Electrical Machines, ICEM2006 Facchinetti A. & Mauri M. (2008). Hardware in the loop test-rig for pantograph active control evaluation. Proceedings of IEEE International Symposium on Industrial Electronics, 2008. ISIE 2008, pp. 2171 – 2176, ISBN 978-1-4244-1665-3, Cambridge, July 2008, IEEE Jian Wu; Yong Cheng; Srivastava A.K.; Schulz N.N.; Ginn, H.L.. (2006). Hardware in the Loop Test for Power System Modeling and Simulation. Proceedings of Power Systems Conference and Exposition, 2006. PSCE '06, pp. 1892-1897, ISBN 1-4244-0177-1, Atlanta GA, November 2006, IEEE PES Kuperman A. & Rabinovici R. (2005) Virtual torque and inertia loading of controlled electric drive. Education, IEEE Transactions on, vol. 48 n.1, pp. 47- 52, ISSN 0018-9359. Maclay D. (1997). Simulation gets into the loop. IEE Review vol. 43 n. 3, pp. 109 – 112, ISSN 0953-5683 Mohan, N.; Robbins, W.P.; Undeland, T.M.; Nilssen, R. & Mo, O. (1994) Simulation of power electronic and motion control systems-an overview. Proceedings of the IEEE,vol. 82 n. 8, pp. 1287 – 1302, ISSN 0018-9219 Mauri M; Castelli Dezza F. & G. Marchegiani (2007). A novel small-scale variable speed hydropower emulator using an inverter-controlled induction motor. Proceedings of European Conference on Power Electronics and Applications, 2007 pp. 1-7, ISBN: 978-92-75815-10-8, Aalborg, September 2007, IEEE Mauri M.; Castelli Dezza F. & Marchegiani G. (2008). Hardware in the Loop (HIL) test bench for small-scale Distributed Generation systems. Proceedings of IEEE International Symposium on Industrial Electronics, 2008. ISIE 2008, pp. 2177 – 2182, ISBN 978-1-4244-1665-3, Cambridge, July 2008, IEEE Shi Y. & Monti A. (2008). FPGA-based fast real time simulation of power systems. Proceedings of Power and Energy Society General Meeting - Conversion and Delivery of Electrical Energy in the 21st Century, pp. 1 – 5, ISBN 978-1-4244-1905-0, July 2008, IEEE Harmonics Reduction Techniques in Renewable Energy Interfacing Converters 153Harmonics Reduction Techniques in Renewable Energy Interfacing ConvertersAli M. Eltamaly, Ph.DX Harmonics Reduction Techniques in Renewable Energy Interfacing Converters Ali M. Eltamaly, Ph.D King Saud University Saudi Arabia 1. Introduction In the recent years there has been growing interest in renewable energy generation along with growing demands for development of a suitable utility interface system. Low harmonics in the line currents of the converters used in the utility interface of the renewable energy system is the highest challenge. Controlled and uncontrolled converters are used widely in the utility interface of renewable energy systems to reduce cost, harmonic contents in line currents and to increase the reliability. These converters suffer from high level of Total Harmonic Distortion (THD) in the utility line currents which creates a lot of problems in the power system. A detailed third harmonic current injection technique has been introduced in details to reduce the harmonic contents in line currents of controlled and uncontrolled converters that used in renewable energy applications. The approach is based on circulating third harmonic current from DC link current to line currents of controlled and uncontrolled converters to reduce its THD. A new injected current shape is computed from analysis to achieve sinusoidal utility line currents. A controllable single switch boost converter connected in shunt is employed to circulate the new injected current shape. A method to implement the proposed approach under varying load condition is shown. Analysis, design, limitations, simulation and experimental result are presented. In modern power electronics converters, three-phase controlled converter is commonly used especially as a rectifier or as an inverter in interfacing renewable energy systems with electric utility (Naik et al., 1995; Mohan, 1992, and Eltamaly, 2008). Three-phase controlled converter has simple construction and control, low cost, and low acoustic noise compared with other solutions like a full-bridge IGBT line side inverters or with Vienna rectifiers. Line current of controlled converter and its FFT components are shown in Fig. 1. It is clear from this waveform that this converter generates high harmonics in the line currents which distort the voltage at the point of common coupling in the power system. A lot of efforts have been performed to reduce harmonic contents in the utility line currents of controlled converter (Pacheco et al., 1994 – Arrillaga et al., 2006). Passive filters have been used in many researches with different configurations (Pacheco et al., 1994 and Bhattacharya et al., 1993), but this technique suffers from bulky, heavy filter elements and sometimes causes resonance problems. Active filters have been used in many researches and it sounds an 9Renewable Energy152 HILS is an important tool to reduce development time and ensure the safety and reliability of DG components. The availability of cost effective tools to perform HILS should lead to an increase in the use of this technique as a standard method to test DG components. 6. References Bouscayrol (2008). Different types of Hardware-In-the-Loop simulation for electric drives. Proceedings of IEEE International Symposium on Industrial Electronics, 2008. ISIE 2008. pp. 2146 – 2151, ISBN 978-1-4244-1665-3, Cambridge, July 2008, IEEE Bin Lu; Xin Wu; Figueroa, H. & Monti, A. (2007). A Low-Cost Real-Time Hardware-in-the-Loop Testing Approach of Power Electronics Controls. Industrial Electronics, IEEE Transactions on, vol. 54 n. 2, pp. 919 – 931, ISSN 0278-0046 Bucca G.; Mapelli F. & Mauri M. (2006). A Real-time environment for industrial electrical drive control board testing. Proceedings of International Conference on Electrical Machines, ICEM2006 Facchinetti A. & Mauri M. (2008). Hardware in the loop test-rig for pantograph active control evaluation. Proceedings of IEEE International Symposium on Industrial Electronics, 2008. ISIE 2008, pp. 2171 – 2176, ISBN 978-1-4244-1665-3, Cambridge, July 2008, IEEE Jian Wu; Yong Cheng; Srivastava A.K.; Schulz N.N.; Ginn, H.L.. (2006). Hardware in the Loop Test for Power System Modeling and Simulation. Proceedings of Power Systems Conference and Exposition, 2006. PSCE '06, pp. 1892-1897, ISBN 1-4244-0177-1, Atlanta GA, November 2006, IEEE PES Kuperman A. & Rabinovici R. (2005) Virtual torque and inertia loading of controlled electric drive. Education, IEEE Transactions on, vol. 48 n.1, pp. 47- 52, ISSN 0018-9359. Maclay D. (1997). Simulation gets into the loop. IEE Review vol. 43 n. 3, pp. 109 – 112, ISSN 0953-5683 Mohan, N.; Robbins, W.P.; Undeland, T.M.; Nilssen, R. & Mo, O. (1994) Simulation of power electronic and motion control systems-an overview. Proceedings of the IEEE,vol. 82 n. 8, pp. 1287 – 1302, ISSN 0018-9219 Mauri M; Castelli Dezza F. & G. Marchegiani (2007). A novel small-scale variable speed hydropower emulator using an inverter-controlled induction motor. Proceedings of European Conference on Power Electronics and Applications, 2007 pp. 1-7, ISBN: 978-92-75815-10-8, Aalborg, September 2007, IEEE Mauri M.; Castelli Dezza F. & Marchegiani G. (2008). Hardware in the Loop (HIL) test bench for small-scale Distributed Generation systems. Proceedings of IEEE International Symposium on Industrial Electronics, 2008. ISIE 2008, pp. 2177 – 2182, ISBN 978-1-4244-1665-3, Cambridge, July 2008, IEEE Shi Y. & Monti A. (2008). FPGA-based fast real time simulation of power systems. Proceedings of Power and Energy Society General Meeting - Conversion and Delivery of Electrical Energy in the 21st Century, pp. 1 – 5, ISBN 978-1-4244-1905-0, July 2008, IEEE Harmonics Reduction Techniques in Renewable Energy Interfacing Converters 153Harmonics Reduction Techniques in Renewable Energy Interfacing ConvertersAli M. Eltamaly, Ph.DX Harmonics Reduction Techniques in Renewable Energy Interfacing Converters Ali M. Eltamaly, Ph.D King Saud University Saudi Arabia 1. Introduction In the recent years there has been growing interest in renewable energy generation along with growing demands for development of a suitable utility interface system. Low harmonics in the line currents of the converters used in the utility interface of the renewable energy system is the highest challenge. Controlled and uncontrolled converters are used widely in the utility interface of renewable energy systems to reduce cost, harmonic contents in line currents and to increase the reliability. These converters suffer from high level of Total Harmonic Distortion (THD) in the utility line currents which creates a lot of problems in the power system. A detailed third harmonic current injection technique has been introduced in details to reduce the harmonic contents in line currents of controlled and uncontrolled converters that used in renewable energy applications. The approach is based on circulating third harmonic current from DC link current to line currents of controlled and uncontrolled converters to reduce its THD. A new injected current shape is computed from analysis to achieve sinusoidal utility line currents. A controllable single switch boost converter connected in shunt is employed to circulate the new injected current shape. A method to implement the proposed approach under varying load condition is shown. Analysis, design, limitations, simulation and experimental result are presented. In modern power electronics converters, three-phase controlled converter is commonly used especially as a rectifier or as an inverter in interfacing renewable energy systems with electric utility (Naik et al., 1995; Mohan, 1992, and Eltamaly, 2008). Three-phase controlled converter has simple construction and control, low cost, and low acoustic noise compared with other solutions like a full-bridge IGBT line side inverters or with Vienna rectifiers. Line current of controlled converter and its FFT components are shown in Fig. 1. It is clear from this waveform that this converter generates high harmonics in the line currents which distort the voltage at the point of common coupling in the power system. A lot of efforts have been performed to reduce harmonic contents in the utility line currents of controlled converter (Pacheco et al., 1994 – Arrillaga et al., 2006). Passive filters have been used in many researches with different configurations (Pacheco et al., 1994 and Bhattacharya et al., 1993), but this technique suffers from bulky, heavy filter elements and sometimes causes resonance problems. Active filters have been used in many researches and it sounds an 9Renewable Energy154int200poIncin conthi199poal.,brasufis haInjmashoconcon 2. FiglineleforsintrabocancurFigeresting option b5 and Lin & Ouwer applicationsreasing number line current but tstruction, and itrd harmonics inj3 and Arrillagawer flow which 1995). Injection nches tuned arofers from bulky very sensitive tormonics reductioection of third hny researches anwn in this resetents of line curverters. Proposed Appure 2 shows thee commutated tctric utility. This fundamental fregle-phase transfnsformer neutralost converter sta be regulated. Trent to be depen. 1. Supply curreut this techniqu, 2004). Hybrid for improving of pulses (Lee ethis technique ne will not be readyection technique et al., 2006). Soincrease the swiof third harmonund third harmconstruction, res the deviation n techniques are armonic to lined (Naik et al., 1arch show the srents and increaroach topology of thehyristor inverte approach contaquency componormer is connec ‘n’. The secondage feeding the dhe duty ratio ofding on the dc linnt and its FFT of e suffers from csolutions using apassive filter pe al., 1997 – Tinsleeds large in size, available from s have been usedme other literattching losses andic current to lineonic frequency (Aonant problems,of L and C valsummarized in ( current using z995 and Pejovicuperiority of thising the power f proposed appror used in interfins a zigzag tranent and very lowted between thery of single-phasc link. By operat the boost convek voltage. controlled conveomplexity and hctive filters andrformance (Bhay, 2003) reduces heavy in weightmanufacturer (Le in (Mohan, 1992ures use switche reduce the syst currents can berrillaga et al., 2 and the current ues. Most of thrMao et al., 19987igzag transform, 2000). These res technique in ractor of the contach to reduce haacing renewablesformer that pre impedance for t dc-link mid-poe transformer is ing single switchrter is varying trter without harmigh cost (Ortega passive filters inttacharya et al., the harmonic co, high in cost, coe, 1998). Early w & Bhattacharyas in the main pem reliability (N achieved by usi006). But, this sin the injection bee phase line c and Singh et al.,er has been shosearches and theeducing the harrolled or uncontrmonics generat energy sourcessents high impehe injection currint ‘o’ and the zconnected to a re the injected curo control the inj onic reduction. et al., high 1997). ntents mplex ork in et al., ass of aik et ng LC cheme ranch urrent 2004). wn in ideas monic rolled ed by with dance ent. A igzag ctifier rent If ection Harmonics Reduction Techniques in Renewable Energy Interfacing Converters 155The optimal rms value and phase angle of injection current to get the minimum THD of supply current is the main issue in this chapter. Previous results say; the best rms value of 3rd harmonic current is equal to the average value of dc link current (Choi et al., 1996 and Lee, 1998. But it is not the only condition required, however, the angle of injection current plays a very important rule in the THD of the line current. So, it is important to determine the relation between the angle of injection current and the firing angle . Fig. 2. The proposed approach For firing angle  =20o (as an example), Fig. 3 (a) shows the utility line current with respect to the voltage of phase ‘a’. In Fig. 3 (b), the 3rd harmonic injection current with respect to phase a voltage. It is clear from this waveform that the injection current leads the phase a voltage by 120o which agree with the analysis and waveforms in the upcoming sections. In Fig. 3(c); the line current with using the 3rd harmonic current along with the phase voltage a. This figure reveals that the angle between If and Ia must be 180o with respect to 3rd harmonic frequency. In Fig. 3(d) the voltage between point ‘d’ and ‘n’’ Vdn, the voltage between point ‘f’ and ‘n’ Vfn and the voltage between ‘o’ and ‘n’ Von along with phase ‘a’ voltage. The third harmonic components of the voltages Vdn, Vfn and Von have been used to inject third harmonic back to utility line current to reduce the harmonic contents in utility line currents. So a careful analysis for these voltages is required to get the optimum value of injection current and its angle. It is convenient to employ Fourier series in the analysis of the distorted waveforms. In general, a non-sinusoidal waveform f(t) can be expressed as follows:     10 sincos)(nnn tnbtnaatf  where a0, an, and bn are Fourier coefficients. By applying Fourier equations to the waveforms of Vdn only to get the third harmonic component as following: Renewable Energy154int200poIncin conthi199poal.,brasufis haInjmashoconcon 2. FiglineleforsintrabocancurFigeresting option b5 and Lin & Ouwer applicationsreasing number line current but tstruction, and itrd harmonics inj3 and Arrillagawer flow which 1995). Injection nches tuned arofers from bulky very sensitive tormonics reductioection of third hny researches anwn in this resetents of line curverters. Proposed Appure 2 shows thee commutated tctric utility. This fundamental fregle-phase transfnsformer neutralost converter sta be regulated. Trent to be depen. 1. Supply curreut this techniqu, 2004). Hybrid for improving of pulses (Lee ethis technique ne will not be readyection technique et al., 2006). Soincrease the swiof third harmonund third harmconstruction, res the deviation n techniques are armonic to lined (Naik et al., 1arch show the srents and increaroach topology of thehyristor inverte approach contaquency componormer is connec ‘n’. The secondage feeding the dhe duty ratio ofding on the dc linnt and its FFT of e suffers from csolutions using apassive filter pe al., 1997 – Tinsleeds large in size, available from s have been usedme other literattching losses andic current to lineonic frequency (Aonant problems,of L and C valsummarized in ( current using z995 and Pejovicuperiority of thising the power f proposed appror used in interfins a zigzag tranent and very lowted between thery of single-phasc link. By operat the boost convek voltage. controlled conveomplexity and hctive filters andrformance (Bhay, 2003) reduces heavy in weightmanufacturer (Le in (Mohan, 1992ures use switche reduce the syst currents can berrillaga et al., 2 and the current ues. Most of thrMao et al., 19987igzag transform, 2000). These res technique in ractor of the contach to reduce haacing renewablesformer that pre impedance for t dc-link mid-poe transformer is ing single switchrter is varying trter without harmigh cost (Ortega passive filters inttacharya et al., the harmonic co, high in cost, coe, 1998). Early w & Bhattacharyas in the main pem reliability (N achieved by usi006). But, this sin the injection bee phase line c and Singh et al.,er has been shosearches and theeducing the harrolled or uncontrmonics generat energy sourcessents high impehe injection currint ‘o’ and the zconnected to a re the injected curo control the inj onic reduction. et al., high 1997). ntents mplex ork in et al., ass of aik et ng LC cheme ranch urrent 2004). wn in ideas monic rolled ed by with dance ent. A igzag ctifier rent If ection Harmonics Reduction Techniques in Renewable Energy Interfacing Converters 155The optimal rms value and phase angle of injection current to get the minimum THD of supply current is the main issue in this chapter. Previous results say; the best rms value of 3rd harmonic current is equal to the average value of dc link current (Choi et al., 1996 and Lee, 1998. But it is not the only condition required, however, the angle of injection current plays a very important rule in the THD of the line current. So, it is important to determine the relation between the angle of injection current and the firing angle . Fig. 2. The proposed approach For firing angle  =20o (as an example), Fig. 3 (a) shows the utility line current with respect to the voltage of phase ‘a’. In Fig. 3 (b), the 3rd harmonic injection current with respect to phase a voltage. It is clear from this waveform that the injection current leads the phase a voltage by 120o which agree with the analysis and waveforms in the upcoming sections. In Fig. 3(c); the line current with using the 3rd harmonic current along with the phase voltage a. This figure reveals that the angle between If and Ia must be 180o with respect to 3rd harmonic frequency. In Fig. 3(d) the voltage between point ‘d’ and ‘n’’ Vdn, the voltage between point ‘f’ and ‘n’ Vfn and the voltage between ‘o’ and ‘n’ Von along with phase ‘a’ voltage. The third harmonic components of the voltages Vdn, Vfn and Von have been used to inject third harmonic back to utility line current to reduce the harmonic contents in utility line currents. So a careful analysis for these voltages is required to get the optimum value of injection current and its angle. It is convenient to employ Fourier series in the analysis of the distorted waveforms. In general, a non-sinusoidal waveform f(t) can be expressed as follows:     10 sincos)(nnn tnbtnaatf  where a0, an, and bn are Fourier coefficients. By applying Fourier equations to the waveforms of Vdn only to get the third harmonic component as following: Renewable Energy156    56363 33 sin * cos 3 2sin 2 sin 48mmVa V t t d t             (1)    56363 33 sin *sin 3 cos 4 2cos 28mmVb V t t d t             (2) From (1) and (2), Vdn3 and its angle can be obtained as in (3) and (4). 2 2 23 3 331 * 1 8sin82LLdnVV a b     (3) where: Vm is the peak value of phase voltage, VLL is the rms value of line to line supply voltage, and, Vdn3 is the rms value of 3rd harmonic of the voltage between points d and n. And angle of 3dnV is       1 1332sin 2 sin 4tan tancos 4 2 cos 2ab                (4) In the same way 3fnV can be obtained as following:    1163763 33 sin *cos3 2sin 2 sin 48mmVa V t t d t             (5)    51163763 33 sin *sin 3 cos 4 2 cos 28mmVb V t t d t             (6) From (5) and (6), 3fnV and its angle as in (7) and (8). 2 2 23 3 331 * 1 8sin82LLfnVV a b     (7) Vfn3 is the rms value of 3rd harmonic of the voltage between points f and n. And angle of 3fnV is;       1 1332sin 2 sin 4tan tancos 4 2 cos 2ab                (8) 3 33 2dn fnonV VV (9) Harmonics Reduction Techniques in Renewable Energy Interfacing Converters 157Then; 233 1 8sin8LLonVV   (10) And angle of Von3, is       1 1332sin 2 sin 4tan tancos 4 2 cos 2ab                 (11) The 3rd harmonic injection current is;  3 3on onfV VIZ Z       (12) where, Von3 is the rms value of 3rd harmonic of the voltage between points o and n, and,  is the angle of Von3, and,  is the angle between Von3 and fI (the impedance angle of injection current path). As shown in Fig. 3, the voltage Va is taken as a reference, so the phase angle of fundamental and 3rd harmonic components of phase a current are - and -3 respectively. As explained in (Eltamaly, 2003) and it is clear from Fig. 3, the optimum phase difference between phase a current and injection current is 180o. Then the difference between the angle of If (which is  ) and the angle of fundamental current which is (- and -3 for fundamental and third harmonic frequencies respectively) is 180o as described in (13).    3 180opt    (13) So the optimum angle of the injection current, opt can be obtained from the following equation;   180 3opt   (14) Also, the optimum angle of the impedance of injection current path, opt can be obtained from the following equation; 3 180opt     (15) From (14) and (15) the phase difference between each vectors for various firing angles;=20o and 40o (as an example for rectifier mode) and  =130o and 150o (as an example for inverter mode) is shown in Fig. 4 (a) and (b) respectively. From (10) the relation between Von3/VLL and the firing angle  is shown in Fig. 5. In the same way; from (14) and (15) the variation of injection current optimum angle, the angle of Von3,  and the angle between If, and Von3 (angle of impedance angle of injection path) along with the firing angle  is shown in Fig. 5. Renewable Energy156    56363 33 sin * cos 3 2sin 2 sin 48mmVa V t t d t             (1)    56363 33 sin *sin 3 cos 4 2cos 28mmVb V t t d t             (2) From (1) and (2), Vdn3 and its angle can be obtained as in (3) and (4). 2 2 23 3 331 * 1 8sin82LLdnVV a b     (3) where: Vm is the peak value of phase voltage, VLL is the rms value of line to line supply voltage, and, Vdn3 is the rms value of 3rd harmonic of the voltage between points d and n. And angle of 3dnV is       1 1332sin 2 sin 4tan tancos 4 2 cos 2ab                (4) In the same way 3fnV can be obtained as following:    1163763 33 sin *cos3 2sin 2 sin 48mmVa V t t d t             (5)    51163763 33 sin *sin 3 cos 4 2 cos 28mmVb V t t d t             (6) From (5) and (6), 3fnV and its angle as in (7) and (8). 2 2 23 3 331 * 1 8sin82LLfnVV a b     (7) Vfn3 is the rms value of 3rd harmonic of the voltage between points f and n. And angle of 3fnV is;       1 1332sin 2 sin 4tan tancos 4 2 cos 2ab                (8) 3 33 2dn fnonV VV (9) Harmonics Reduction Techniques in Renewable Energy Interfacing Converters 157Then; 233 1 8sin8LLonVV   (10) And angle of Von3, is       1 1332sin 2 sin 4tan tancos 4 2 cos 2ab                 (11) The 3rd harmonic injection current is;  3 3on onfV VIZ Z       (12) where, Von3 is the rms value of 3rd harmonic of the voltage between points o and n, and,  is the angle of Von3, and,  is the angle between Von3 and fI (the impedance angle of injection current path). As shown in Fig. 3, the voltage Va is taken as a reference, so the phase angle of fundamental and 3rd harmonic components of phase a current are - and -3 respectively. As explained in (Eltamaly, 2003) and it is clear from Fig. 3, the optimum phase difference between phase a current and injection current is 180o. Then the difference between the angle of If (which is  ) and the angle of fundamental current which is (- and -3 for fundamental and third harmonic frequencies respectively) is 180o as described in (13).    3 180opt    (13) So the optimum angle of the injection current, opt can be obtained from the following equation;   180 3opt   (14) Also, the optimum angle of the impedance of injection current path, opt can be obtained from the following equation; 3 180opt     (15) From (14) and (15) the phase difference between each vectors for various firing angles;=20o and 40o (as an example for rectifier mode) and  =130o and 150o (as an example for inverter mode) is shown in Fig. 4 (a) and (b) respectively. From (10) the relation between Von3/VLL and the firing angle  is shown in Fig. 5. In the same way; from (14) and (15) the variation of injection current optimum angle, the angle of Von3,  and the angle between If, and Von3 (angle of impedance angle of injection path) along with the firing angle  is shown in Fig. 5. Renewable Energy158 (a) Voltage and current of phase a. (b) Voltage of phase a and reinjection current. (c) Voltage, current of phase a and reinjection current. (d) The voltages Vdn, Von, and Vfn. Fig. 3. Various voltages and currents of the proposed approach at =20o. 3. Design Example Experimental and simulation verification for this technique has been carried out. The prototype model is used in simulation and in the experiment. DC link is connected to three-phase controlled converter and electrical utility as shown in Fig. 2. For the controlled rectifier, the values of 3onV varies between Von3(=0)=0.1194pu to Von3(=90o)=0.358pu. At full load the dc current Io= 0.6224 pu. The base voltage and current are 220V and 3.3A, respectively. The simulation results in the next section reveals that the minimum THD occurs almost at If/Io=1.15, so at full load; If=0.716pu. The rated third harmonic current passing through zigzag transformer is If /3=0.239pu. The value of If is controlled depending on the value of dc link current for minimum THD. The third harmonic injection current can be controlled by controlling the duty ratio of the boost converter. Current sensor is used to measure the Harmonics Reduction Techniques in Renewable Energy Interfacing Converters 159actual dc link current and the 3rd harmonic injection current. The error signal between these two currents is used to control the duty ratio of the boost converter. 10 kHz switching frequency of the boost converter is used in simulation and in the practical prototype. The schematic of the controlled circuit of the boost converter is shown in Fig. 6. Design of Single-Phase Transformer The rated primary voltage of single phase transformer in 3rd harmonic injection pass is   puVon 358.0903  . Also the rated primary current of this transformer is puI f 716.0 . So, the rated kVA of this transformer is pu256.0358.0*716.0  . The turns ratio of step up single-phase must handle the minimum value as well as the maximum value of 3onV . The minimum value of 3onV is at 0 which is   1194.003 onV pu. Then, the corresponding value of 3onV  is:  3 0 0.1194 /onV n   pu (16) The output dc voltage of single-phase diode rectifier, biV is: 2 * 2 * /bi onV V  (17) The minimum value of 3onV is corresponding on the maximum value of the duty ratio of boost converter. Assume the maximum allowable value of the duty ratio is 0.8. Then the output voltage of the boost converter, boV is:  / 1bo biV V D  (18) Substitute the value of biV from (17) into (18) and D=0.8, the following equation can be obtained:        0.8 2 * 2 * / / 1 0.8 10 * 2 * /bo on onV D V V      (19) The output voltage from the boost converter must equal the dc link voltage. The dc link voltage is kept constant at 1.15 pu by controlling the firing angle of the controlled converter and the modulation index of PWM inverter. So, by equating (19) with 2.2pu, the following equation can be obtained: puVon 2.2/*2*10 3   , Then;    3 0 2.2 * / 10 2onV pu    (20) By equating the values of 3onV  in (14) and (18), the following equation can be obtained:  210/*2.2/1194.0 n Then the turns ratio of the single-phase transformer is:  10 2 * 0.1194 / 2.2 * 0.24n   (21) Renewable Energy158 (a) Voltage and current of phase a. (b) Voltage of phase a and reinjection current. (c) Voltage, current of phase a and reinjection current. (d) The voltages Vdn, Von, and Vfn. Fig. 3. Various voltages and currents of the proposed approach at =20o. 3. Design Example Experimental and simulation verification for this technique has been carried out. The prototype model is used in simulation and in the experiment. DC link is connected to three-phase controlled converter and electrical utility as shown in Fig. 2. For the controlled rectifier, the values of 3onV varies between Von3(=0)=0.1194pu to Von3(=90o)=0.358pu. At full load the dc current Io= 0.6224 pu. The base voltage and current are 220V and 3.3A, respectively. The simulation results in the next section reveals that the minimum THD occurs almost at If/Io=1.15, so at full load; If=0.716pu. The rated third harmonic current passing through zigzag transformer is If /3=0.239pu. The value of If is controlled depending on the value of dc link current for minimum THD. The third harmonic injection current can be controlled by controlling the duty ratio of the boost converter. Current sensor is used to measure the Harmonics Reduction Techniques in Renewable Energy Interfacing Converters 159actual dc link current and the 3rd harmonic injection current. The error signal between these two currents is used to control the duty ratio of the boost converter. 10 kHz switching frequency of the boost converter is used in simulation and in the practical prototype. The schematic of the controlled circuit of the boost converter is shown in Fig. 6. Design of Single-Phase Transformer The rated primary voltage of single phase transformer in 3rd harmonic injection pass is   puVon 358.0903  . Also the rated primary current of this transformer is puI f 716.0 . So, the rated kVA of this transformer is pu256.0358.0*716.0  . The turns ratio of step up single-phase must handle the minimum value as well as the maximum value of 3onV . The minimum value of 3onV is at 0 which is   1194.003 onV pu. Then, the corresponding value of 3onV  is:  3 0 0.1194 /onV n   pu (16) The output dc voltage of single-phase diode rectifier, biV is: 2 * 2 * /bi onV V  (17) The minimum value of 3onV is corresponding on the maximum value of the duty ratio of boost converter. Assume the maximum allowable value of the duty ratio is 0.8. Then the output voltage of the boost converter, boV is:  / 1bo biV V D  (18) Substitute the value of biV from (17) into (18) and D=0.8, the following equation can be obtained:        0.8 2 * 2 * / / 1 0.8 10 * 2 * /bo on onV D V V      (19) The output voltage from the boost converter must equal the dc link voltage. The dc link voltage is kept constant at 1.15 pu by controlling the firing angle of the controlled converter and the modulation index of PWM inverter. So, by equating (19) with 2.2pu, the following equation can be obtained: puVon 2.2/*2*10 3   , Then;    3 0 2.2 * / 10 2onV pu    (20) By equating the values of 3onV  in (14) and (18), the following equation can be obtained:  210/*2.2/1194.0 n Then the turns ratio of the single-phase transformer is:  10 2 * 0.1194 / 2.2 * 0.24n   (21) Renewable Energy160aV3onVfIopt3aI3(a) o20opt(b) aV3onVfIopt3aI3o40opt (a) Rectifier aV3onVfIopt3aI3o130optaV3onVfIopt3aI3o150opt (b) Inverter Fig. 4. Phase difference between each component in a rectifier and inverter Harmonics Reduction Techniques in Renewable Energy Interfacing Converters 161FigFig V/V. 5. The variation. 6. The schemat000.10.20.30.4V on3 /VLL0-200-1000100200angle oIfI of Von3, , opt, ic of the control c20 40 620 40 6opt along with  ircuit of the boos0 80 100 80 10 , D. t converter. 0 120 1400 120 140egrees 160 180160 180   Renewable Energy160aV3onVfIopt3aI3(a) o20opt(b) aV3onVfIopt3aI3o40opt (a) Rectifier aV3onVfIopt3aI3o130optaV3onVfIopt3aI3o150opt (b) Inverter Fig. 4. Phase difference between each component in a rectifier and inverter Harmonics Reduction Techniques in Renewable Energy Interfacing Converters 161FigFig V/V. 5. The variation. 6. The schemat000.10.20.30.4V on3 /VLL0-200-1000100200angle oIfI of Von3, , opt, ic of the control c20 40 620 40 6opt along with  ircuit of the boos0 80 100 80 10 , D. t converter. 0 120 1400 120 140egrees 160 180160 180   Renewable Energy1624. Simulation and Experimental Results The simulation of the proposed technique has been performed by using PSIM computer program (PSIM6.1). Same values of the components used in simulation program have been used in the experimental prototype to compare these results. Fig. 7 shows the relation between THD and the value of If /Io for different values of firing angle, . This figure reveals that the optimal value of If /Io is about 1.15 for the best THD. Also, This figure shows the importance of 3rd harmonic injection technique especially for high value of firing angle, . Fig. 8 shows the relation between THD and the angle of If with respect to Von,  at If /Io=1.15 for different values of firing angle, . It is clear that the best THD occurs around opt which obtained from (13). The value of opt increase with increasing the firing angle . The simulation and experimental waveforms are shown for =20o and 40o as an examples. 4.1 Simulation and experimental results at =20o Figure 9 shows the waveforms of voltage Vdf and Vab. Fig. 10 shows the supply current waveform and its FFT components with respect to phase a voltage without injection of 3rd harmonic current. It is clear from this figure that the supply current has very high THD mostly of 5th and 7th harmonics. Fig. 11 shows the voltage of Vdn, and Vfn. It is shown in the top of the experimental result waveform that the frequency of this voltage is 180o which is the 3rd harmonic voltage and the value of rms voltage is 39.9 that can be obtained from (10) at =20o. Figure 12 shows the supply current waveform and its FFT components with respect to phase a voltage with optimum amplitude and angle of 3rd harmonic current injection. It is clear from this figure that the supply current becomes very near to sine-wave with 5% THD. Figure 13 shows supply current waveform Ia and optimum value and angle of injection 3rd harmonic current with respect to phase a voltage. It is clear from these waveforms that the value of the angle between If and Va is about 120o and the angle between Ia and If is 180o which agree with the vector diagram shown in Fig. 4 (a). Harmonics Reduction Techniques in Renewable Energy Interfacing Converters 163 Fig. 7. The relation between THD and the value of If /Io. Fig. 8. The relation between THD and the angle of If with respect to Von at If /Io=1.15 for different values of firing angle, . 0.5 1 1.5 201020304050If /IoTHD (%)0 20 40 60 80010203040506070THD% Renewable Energy1624. Simulation and Experimental Results The simulation of the proposed technique has been performed by using PSIM computer program (PSIM6.1). Same values of the components used in simulation program have been used in the experimental prototype to compare these results. Fig. 7 shows the relation between THD and the value of If /Io for different values of firing angle, . This figure reveals that the optimal value of If /Io is about 1.15 for the best THD. Also, This figure shows the importance of 3rd harmonic injection technique especially for high value of firing angle, . Fig. 8 shows the relation between THD and the angle of If with respect to Von,  at If /Io=1.15 for different values of firing angle, . It is clear that the best THD occurs around opt which obtained from (13). The value of opt increase with increasing the firing angle . The simulation and experimental waveforms are shown for =20o and 40o as an examples. 4.1 Simulation and experimental results at =20o Figure 9 shows the waveforms of voltage Vdf and Vab. Fig. 10 shows the supply current waveform and its FFT components with respect to phase a voltage without injection of 3rd harmonic current. It is clear from this figure that the supply current has very high THD mostly of 5th and 7th harmonics. Fig. 11 shows the voltage of Vdn, and Vfn. It is shown in the top of the experimental result waveform that the frequency of this voltage is 180o which is the 3rd harmonic voltage and the value of rms voltage is 39.9 that can be obtained from (10) at =20o. Figure 12 shows the supply current waveform and its FFT components with respect to phase a voltage with optimum amplitude and angle of 3rd harmonic current injection. It is clear from this figure that the supply current becomes very near to sine-wave with 5% THD. Figure 13 shows supply current waveform Ia and optimum value and angle of injection 3rd harmonic current with respect to phase a voltage. It is clear from these waveforms that the value of the angle between If and Va is about 120o and the angle between Ia and If is 180o which agree with the vector diagram shown in Fig. 4 (a). Harmonics Reduction Techniques in Renewable Energy Interfacing Converters 163 Fig. 7. The relation between THD and the value of If /Io. Fig. 8. The relation between THD and the angle of If with respect to Von at If /Io=1.15 for different values of firing angle, . 0.5 1 1.5 201020304050If /IoTHD (%)0 20 40 60 80010203040506070THD% Renewable Energy164FigFigvo Fig. 9. The wavefor(a) Expe. 10. The supplltage without inje. 11. The voltagems of voltage Vdfrimental result. y current wavefction of 3rd harm of Vdn, and Vfn and Vab. orm and its FFTonic current. (b) FFT comp components w onents of Ia ith respect to ph ase a Harmonics Reduction Techniques in Renewable Energy Interfacing Converters 165Figvo. 12. The supplltage at optimum(b(c)y current wavef 3rd harmonic cu(a) Simulation res ) Experimental r FFT componentorm and its FFTrrent injection. ult. esult. s of Ia. components w ith respect to phase a Renewable Energy164FigFigvo Fig. 9. The wavefor(a) Expe. 10. The supplltage without inje. 11. The voltagems of voltage Vdfrimental result. y current wavefction of 3rd harm of Vdn, and Vfn and Vab. orm and its FFTonic current. (b) FFT comp components w onents of Ia ith respect to ph ase a Harmonics Reduction Techniques in Renewable Energy Interfacing Converters 165Figvo. 12. The supplltage at optimum(b(c)y current wavef 3rd harmonic cu(a) Simulation res ) Experimental r FFT componentorm and its FFTrrent injection. ult. esult. s of Ia. components w ith respect to phase a Renewable Energy166 (a) Simulation result for Va and If. (b) Experimental result for Va and If. (c) Experimental result for Ia and If Fig. 13. The supply current waveform Ia and optimum value and angle of injection 3rd harmonic current with respect to phase a voltage. Harmonics Reduction Techniques in Renewable Energy Interfacing Converters 1674.2 Simulation and experimental results at =40o Figure 14 shows the waveforms of voltage Vdf and Vab. Fig. 15 shows the supply current waveform and its FFT components along with phase a voltage without injection of 3rd harmonic current. It is clear from this figure that the supply current has very high THD mostly of 5th and 7th harmonics. It is clear that this THD is high than the one of =40o. Figure 16 shows the voltage of Vdn, and Vfn and their FFT components with respect to phase a voltage. It is clear from this figure that these voltages have triplex harmonics and the 3rd harmonic is the most dominant harmonic. It is also clear that both voltages Vdn, and Vfn have the same harmonics. Figure 17 shows supply current waveform and its FFT components with optimum value and angle of injection 3rd harmonic current. It is clear from this figure that the supply current becomes very near to sine-wave with 6% THD. Figure 18 shows the supply current waveform Ia and optimum value and angle of 3rd harmonic injection current with respect to phase a voltage. It is clear from these waveforms that the angle between If and Va is about 60o which agree with the vector diagram in Fig. 4 (b). Fig. 14. The waveforms of voltage Vdf and Vab. Renewable Energy166 (a) Simulation result for Va and If. (b) Experimental result for Va and If. (c) Experimental result for Ia and If Fig. 13. The supply current waveform Ia and optimum value and angle of injection 3rd harmonic current with respect to phase a voltage. Harmonics Reduction Techniques in Renewable Energy Interfacing Converters 1674.2 Simulation and experimental results at =40o Figure 14 shows the waveforms of voltage Vdf and Vab. Fig. 15 shows the supply current waveform and its FFT components along with phase a voltage without injection of 3rd harmonic current. It is clear from this figure that the supply current has very high THD mostly of 5th and 7th harmonics. It is clear that this THD is high than the one of =40o. Figure 16 shows the voltage of Vdn, and Vfn and their FFT components with respect to phase a voltage. It is clear from this figure that these voltages have triplex harmonics and the 3rd harmonic is the most dominant harmonic. It is also clear that both voltages Vdn, and Vfn have the same harmonics. Figure 17 shows supply current waveform and its FFT components with optimum value and angle of injection 3rd harmonic current. It is clear from this figure that the supply current becomes very near to sine-wave with 6% THD. Figure 18 shows the supply current waveform Ia and optimum value and angle of 3rd harmonic injection current with respect to phase a voltage. It is clear from these waveforms that the angle between If and Va is about 60o which agree with the vector diagram in Fig. 4 (b). Fig. 14. The waveforms of voltage Vdf and Vab. Renewable Energy168FigwiFig(a) Expe. 15. The supplythout injection of . 16. The voltage rimental result. current wavefor 3rd harmonic cur (a(b)of Vdn, and Vfn an m and its FFT crent. ) Experimental r FFT componentd their FFT comp(b) FFT compomponents alongesult. s of Ia. onents with responents of Ia with phase a v ect to phase a volt oltage age. Harmonics Reduction Techniques in Renewable Energy Interfacing Converters 169Figan Figha. 17. The supplygle of injection 3r(a) Simu. 18. The supplrmonic current w(a(b) current wavefod harmonic currelation result. y current wavefith respect to ph) Experimental r FFT componentrm and its FFT nt. orm Ia and optiase a voltage. esult. s of Ia. components wit(b) Experimemum value and h optimum valuntal result. angle of injectie and on 3rd Renewable Energy168FigwiFig(a) Expe. 15. The supplythout injection of . 16. The voltage rimental result. current wavefor 3rd harmonic cur (a(b)of Vdn, and Vfn an m and its FFT crent. ) Experimental r FFT componentd their FFT comp(b) FFT compomponents alongesult. s of Ia. onents with responents of Ia with phase a v ect to phase a volt oltage age. Harmonics Reduction Techniques in Renewable Energy Interfacing Converters 169Figan Figha. 17. The supplygle of injection 3r(a) Simu. 18. The supplrmonic current w(a(b) current wavefod harmonic currelation result. y current wavefith respect to ph) Experimental r FFT componentrm and its FFT nt. orm Ia and optiase a voltage. esult. s of Ia. components wit(b) Experimemum value and h optimum valuntal result. angle of injectie and on 3rd Renewable Energy1704. Conclusions Third harmonics injection technique plays a significant rule in reducing of the THD of the utility line current of three-phase controlled converters. The optimal amplitude and phase angle of injection current slightly changes with the firing angle. In this research, a detailed analysis for determining the relation between the optimal amplitude and phase angle of injection current along with the firing angle has been carried out. The THD of line current is highly affected by the amplitude of 3rd harmonic injection current as well as its phase angle with supply voltage. The optimum amplitude and angle of 3rd harmonic current have been determined by a detailed mathematical analysis of the system. The THD of the utility line current from simulation and the experimental results proves the mathematical results for this technique. The THD of the utility line current with optimal harmonic injection current is lower than limits of harmonics standards. 5. References A. Ametani "Harmonic reduction in thyristor converters by harmonic current injection" IEEE Trans. on Power Apparatus and Systems, vol. PAS-95, no. 2 , March/April 1976, pp.441-450. Ali M. Eltamaly "A Modified Harmonics Reduction Technique of Three-Phase Controlled Converter" IEEE Trans. Industrial Electronics, vol.55, No.3 March. 2008, pp. 1190-1198. Ali M. Eltamaly" A new relation between firing angle of three-phase controlled converter and best angle of injection current " in IEEE Conf. Proc. MEPCON’2003, Minoufiya University, Shebin El-Kom, Egypt, Dec. 2003, pp. 793-798. B. Lin, Y. Ou "Active power filter based on three-phase two-leg switch-clamped inverter" Electric Power Systems Research, vol. 72, June 2004, pp. 63–72. B. S. Lee “New clean power reactor systems for utility interface of static converters”, Ph.D. Thesis Texas A&M University, Aug. 1998. B. S. Lee; P. N. Enjeti, I. J. Pitel “An optimized active interphase transformer for auto-connected 12-pulse rectifiers results in clean input power” in IEEE Conf. Proc. APEC1997, Atlanta, Georgia, USA, vol. 2, Feb. 1997, pp. 666-671. Bhim Singh, Brij N. Singh, Ambrish Chandra, Kamal Al-Haddad, Ashish Pandey, Dwarka P. Kothari, "A review of three-phase improved power quality ac–dc converters" IEEE Trans. on Industrial Electronics, vol. 51, no. 3, June 2004. C. T. Tinsley" Modeling of multi-pulse transformer rectifier units in power distribution systems" MS.c., Thesis Virginia Polytechnic Institute, USA, Aug. 2003. Cassiano Rech, and José R. Pinheiro, "Line current harmonics reduction in multipulse connection of asymmetrically loaded rectifiers" IEEE Trans. on Industrial Electronics, vol. 52, no. 3, June 2005. H. Mao, F. C. Y. Lee, D. Boroyevich, and S. Hiti, "Review of high-performance three-phase power-factor correction circuits" IEEE Trans. on Industrial Electronics, vol. 44, no.4, Aug. 1997, pp. 437-446. H.A. Pacheco; G. Jimenez, J. Arau "Optimization method to design filters for harmonic current reduction in a three phase rectifier" in IEEE Conf. proc. CIEP'94, Puebla, Mexico, Aug. 1994, pp. 138-144. Harmonics Reduction Techniques in Renewable Energy Interfacing Converters 171J. Marafao, J. Pomilio, and G. Spiazz " Improved three-phase high-quality rectifier with line-commutated switches" IEEE Trans. on Power Electronics, vol.19, no.3, May 2004, pp. 640-648. J. Ortega, M. Esteve, M. Payán, A. Gómez "Reference current computation methods for active power filters: accuracy assessment in the frequency domain" IEEE Trans. on Power Electronics, vol. 20, no. 2, March 2005, pp. 446-456. Jos Arrillaga, Y. H. Liu, Lasantha B. Perera, and Neville R. Watson "A current reinjection scheme that adds self-commutation and pulse multiplication to the thyristor converter" IEEE Trans. on Power Delivery, vol.21, no. 3, July 2006. M. Rastogi, N. Mohan, and C. P. Henze " Three-phase sinusoidal current rectifier with zero-current switching" IEEE Trans. on Power Electronics, vol. 10, no. 6, Nov. 1995, pp. 753-759. N. Mohan “A novel Approach to minimize line current harmonics in interfacing renewable energy sources with 3-phase utility systems” in IEEE Conf. Proc. APEC'92, Boston Massachusetts, Feb. 1992, pp. 852 –858. N. Mohan, M. Rastogi and R. Naik " Analysis of a new power electronics interface with approximately sinusoidal 3-phase utility currents and a regulated dc output" IEEE Trans. on Power Delivery, April 1993, pp. 540-546. P. Pejovic, D. Shmilovitz " Low-harmonic thyristor rectifiers applying current injection" IEEE TRANS. on Aerospace and Electronic Systems vol. 39, no. 4, Oct. 2003, pp.1365-1374. P. Pejovic, P. Bojzovic, and D. Shmilovitz "Low-harmonic, three-phase rectifier that applies current injection and a passive resistance emulator" IEEE Power Electronics Letters, vol. 3, no. 3, Sept. 2005, pp. 96-100. P. Pejovic" Two three-phase high power factor rectifiers that apply the third harmonic current injection and passive resistance emulation", IEEE Trans. on Power Electronics, vol. 15, no. 6, Nov. 2000, pp. 1228-1240. PSIM6.1, available: www.powersimtech.com. R. Naik, N. Mohan, M. Rogers and A. Bulawka” A novel grid interface, optimized for utility-scale applications of photovoltaic, wind-electric, and fuel-cell systems” IEEE Trns. on Power Delivery, vol.10, no.4, Oct. 1995, pp.1920-1926. S. Bhattacharya, D. M. Divan; B. B. Banerjee; "Control and reduction of terminal voltage harmonic distortion (THD) in hybrid series active and parallel passive filter system" in IEEE Conf. Proc. PESC 93, Seatle, USA, June, 1993, pp. 779-786. S. Bhattacharya, P. Cheng, and D. M. Divan, " Hybrid solutions for improving passive filter performance in high power applications" IEEE Trans. on Industry Applications, vol. 33, no. 3, May/June 1997, pp.732-747. S. Choi, P. N. Enjeti, H. Lee, and I. I. Pital “ A new active interface reactor 12-pulse rectifiers provides clean power utility interface “ IEEE Trans. on Industry Applications, vol. 32, No. 6, 1996, pp. 1304-1311. S. Pal "Simulation of current mode control schemes for power factor correction circuits" MS.c thesis, Faculty of Engineering and Applied Science, Mernorial University of Newfoundland, St. John's Newfoundland Canada, May 1998. W. Mielczarski, W. B. Lawrance, R. Nowacki, and D. G. Holmes " Harmonic current reduction in three-phase bridge-rectifier circuits using controlled current injection" IEEE Trans. on Industrial Electronics, vol. 44, no. 5, Oct. 1997, pp. 604-611. Renewable Energy1704. Conclusions Third harmonics injection technique plays a significant rule in reducing of the THD of the utility line current of three-phase controlled converters. The optimal amplitude and phase angle of injection current slightly changes with the firing angle. In this research, a detailed analysis for determining the relation between the optimal amplitude and phase angle of injection current along with the firing angle has been carried out. The THD of line current is highly affected by the amplitude of 3rd harmonic injection current as well as its phase angle with supply voltage. The optimum amplitude and angle of 3rd harmonic current have been determined by a detailed mathematical analysis of the system. The THD of the utility line current from simulation and the experimental results proves the mathematical results for this technique. The THD of the utility line current with optimal harmonic injection current is lower than limits of harmonics standards. 5. References A. Ametani "Harmonic reduction in thyristor converters by harmonic current injection" IEEE Trans. on Power Apparatus and Systems, vol. PAS-95, no. 2 , March/April 1976, pp.441-450. Ali M. Eltamaly "A Modified Harmonics Reduction Technique of Three-Phase Controlled Converter" IEEE Trans. Industrial Electronics, vol.55, No.3 March. 2008, pp. 1190-1198. Ali M. Eltamaly" A new relation between firing angle of three-phase controlled converter and best angle of injection current " in IEEE Conf. Proc. MEPCON’2003, Minoufiya University, Shebin El-Kom, Egypt, Dec. 2003, pp. 793-798. B. Lin, Y. Ou "Active power filter based on three-phase two-leg switch-clamped inverter" Electric Power Systems Research, vol. 72, June 2004, pp. 63–72. B. S. Lee “New clean power reactor systems for utility interface of static converters”, Ph.D. Thesis Texas A&M University, Aug. 1998. B. S. Lee; P. N. Enjeti, I. J. Pitel “An optimized active interphase transformer for auto-connected 12-pulse rectifiers results in clean input power” in IEEE Conf. Proc. APEC1997, Atlanta, Georgia, USA, vol. 2, Feb. 1997, pp. 666-671. Bhim Singh, Brij N. Singh, Ambrish Chandra, Kamal Al-Haddad, Ashish Pandey, Dwarka P. Kothari, "A review of three-phase improved power quality ac–dc converters" IEEE Trans. on Industrial Electronics, vol. 51, no. 3, June 2004. C. T. Tinsley" Modeling of multi-pulse transformer rectifier units in power distribution systems" MS.c., Thesis Virginia Polytechnic Institute, USA, Aug. 2003. Cassiano Rech, and José R. Pinheiro, "Line current harmonics reduction in multipulse connection of asymmetrically loaded rectifiers" IEEE Trans. on Industrial Electronics, vol. 52, no. 3, June 2005. H. Mao, F. C. Y. Lee, D. Boroyevich, and S. Hiti, "Review of high-performance three-phase power-factor correction circuits" IEEE Trans. on Industrial Electronics, vol. 44, no.4, Aug. 1997, pp. 437-446. H.A. Pacheco; G. Jimenez, J. Arau "Optimization method to design filters for harmonic current reduction in a three phase rectifier" in IEEE Conf. proc. CIEP'94, Puebla, Mexico, Aug. 1994, pp. 138-144. Harmonics Reduction Techniques in Renewable Energy Interfacing Converters 171J. Marafao, J. Pomilio, and G. Spiazz " Improved three-phase high-quality rectifier with line-commutated switches" IEEE Trans. on Power Electronics, vol.19, no.3, May 2004, pp. 640-648. J. Ortega, M. Esteve, M. Payán, A. Gómez "Reference current computation methods for active power filters: accuracy assessment in the frequency domain" IEEE Trans. on Power Electronics, vol. 20, no. 2, March 2005, pp. 446-456. Jos Arrillaga, Y. H. Liu, Lasantha B. Perera, and Neville R. Watson "A current reinjection scheme that adds self-commutation and pulse multiplication to the thyristor converter" IEEE Trans. on Power Delivery, vol.21, no. 3, July 2006. M. Rastogi, N. Mohan, and C. P. Henze " Three-phase sinusoidal current rectifier with zero-current switching" IEEE Trans. on Power Electronics, vol. 10, no. 6, Nov. 1995, pp. 753-759. N. Mohan “A novel Approach to minimize line current harmonics in interfacing renewable energy sources with 3-phase utility systems” in IEEE Conf. Proc. APEC'92, Boston Massachusetts, Feb. 1992, pp. 852 –858. N. Mohan, M. Rastogi and R. Naik " Analysis of a new power electronics interface with approximately sinusoidal 3-phase utility currents and a regulated dc output" IEEE Trans. on Power Delivery, April 1993, pp. 540-546. P. Pejovic, D. Shmilovitz " Low-harmonic thyristor rectifiers applying current injection" IEEE TRANS. on Aerospace and Electronic Systems vol. 39, no. 4, Oct. 2003, pp.1365-1374. P. Pejovic, P. Bojzovic, and D. Shmilovitz "Low-harmonic, three-phase rectifier that applies current injection and a passive resistance emulator" IEEE Power Electronics Letters, vol. 3, no. 3, Sept. 2005, pp. 96-100. P. Pejovic" Two three-phase high power factor rectifiers that apply the third harmonic current injection and passive resistance emulation", IEEE Trans. on Power Electronics, vol. 15, no. 6, Nov. 2000, pp. 1228-1240. PSIM6.1, available: www.powersimtech.com. R. Naik, N. Mohan, M. Rogers and A. Bulawka” A novel grid interface, optimized for utility-scale applications of photovoltaic, wind-electric, and fuel-cell systems” IEEE Trns. on Power Delivery, vol.10, no.4, Oct. 1995, pp.1920-1926. S. Bhattacharya, D. M. Divan; B. B. Banerjee; "Control and reduction of terminal voltage harmonic distortion (THD) in hybrid series active and parallel passive filter system" in IEEE Conf. Proc. PESC 93, Seatle, USA, June, 1993, pp. 779-786. S. Bhattacharya, P. Cheng, and D. M. Divan, " Hybrid solutions for improving passive filter performance in high power applications" IEEE Trans. on Industry Applications, vol. 33, no. 3, May/June 1997, pp.732-747. S. Choi, P. N. Enjeti, H. Lee, and I. I. Pital “ A new active interface reactor 12-pulse rectifiers provides clean power utility interface “ IEEE Trans. on Industry Applications, vol. 32, No. 6, 1996, pp. 1304-1311. S. Pal "Simulation of current mode control schemes for power factor correction circuits" MS.c thesis, Faculty of Engineering and Applied Science, Mernorial University of Newfoundland, St. John's Newfoundland Canada, May 1998. W. Mielczarski, W. B. Lawrance, R. Nowacki, and D. G. Holmes " Harmonic current reduction in three-phase bridge-rectifier circuits using controlled current injection" IEEE Trans. on Industrial Electronics, vol. 44, no. 5, Oct. 1997, pp. 604-611. Hybrid Control of DC-DC Power Converters 173Hybrid Control of DC-DC Power ConvertersIlse Cervantes, Francisco J. Perez-Pinal and Angelica Mendoza-TorresX Hybrid Control of DC-DC Power Converters Ilse Cervantes*, Francisco J. Perez-Pinal† and Angelica Mendoza-Torres**Applied Mathematics Division. Institute for Scientific andTechnological Research of San Luis Potosi (IPICYT) †National Polytechnic Institute, IPNMexico 1. Introduction. Review of Renewable EnergyIn recent years, there has been an increasing interest in applying renewable energy (RE) in electricity generation and transportation. In accordance to Enerdata, an independent energy consulting and information services company (Enerdata, 2008a), the global electricityneeds in 2007 were closed to 19,900 TWh (almost 40 % of the overall energy consumption),where just 18 % of the total electricity production was coped by some source of RE. Additionally, it is estimated that the electricity demand will grow up to 60 % of theworldwide energy in 2040 (Enerdata, 2008b). This means that, in order to increase energy production almost 50 % in 30 years, some energy generation and distribution problems mustbe solved to obtain feasible and reliable applications. For example, the concept of distributedgeneration (DG) has to be revisited in view of the existence of multiple and highly disperseenergy sources; as well as further investigation about High Voltage Direct Current (HVDC)energy transmission, Flexible AC Transmission Systems (FACTS) and Micro grids. Furthermore, power electronics can account for, as much as 30–40 % of a distributedgeneration facility cost (ECPE et al., 2007). This is due to the fact that most utility scale power electronics are custom designed, which generates an expensive installation. That is,there are also important concerns such as price reduction and system reliability that need further investigation. On the other hand, transportation demands about 28 % of the worldwide energy and it is expected an increment up to 40% in 2040 (International Energy Outlook 2008). In order toreduce this tendency, during the past three decades, a rapid development of more efficientvehicles has been performed. Those efforts have been focused on two main areas: i) thedevelopment of high efficient and high-density power converters, and ii) the improvementof robust and flexible modern control techniques. However, in spite of some advances, thereare still open questions regarding the achievement of three of the most important issues in RE applications: reliability, efficiency, and cost.10Renewable Energy172 Hybrid Control of DC-DC Power Converters 173Hybrid Control of DC-DC Power ConvertersIlse Cervantes, Francisco J. Perez-Pinal and Angelica Mendoza-TorresX Hybrid Control of DC-DC Power Converters Ilse Cervantes*, Francisco J. Perez-Pinal† and Angelica Mendoza-Torres* *Applied Mathematics Division. Institute for Scientific andTechnological Research of San Luis Potosi (IPICYT) †National Polytechnic Institute, IPN Mexico 1. Introduction. Review of Renewable EnergyIn recent years, there has been an increasing interest in applying renewable energy (RE) in electricity generation and transportation. In accordance to Enerdata, an independent energy consulting and information services company (Enerdata, 2008a), the global electricity needs in 2007 were closed to 19,900 TWh (almost 40 % of the overall energy consumption), where just 18 % of the total electricity production was coped by some source of RE. Additionally, it is estimated that the electricity demand will grow up to 60 % of the worldwide energy in 2040 (Enerdata, 2008b). This means that, in order to increase energy production almost 50 % in 30 years, some energy generation and distribution problems must be solved to obtain feasible and reliable applications. For example, the concept of distributed generation (DG) has to be revisited in view of the existence of multiple and highly disperse energy sources; as well as further investigation about High Voltage Direct Current (HVDC) energy transmission, Flexible AC Transmission Systems (FACTS) and Micro grids. Furthermore, power electronics can account for, as much as 30–40 % of a distributed generation facility cost (ECPE et al., 2007). This is due to the fact that most utility scale power electronics are custom designed, which generates an expensive installation. That is, there are also important concerns such as price reduction and system reliability that need further investigation. On the other hand, transportation demands about 28 % of the worldwide energy and it is expected an increment up to 40% in 2040 (International Energy Outlook 2008). In order to reduce this tendency, during the past three decades, a rapid development of more efficient vehicles has been performed. Those efforts have been focused on two main areas: i) the development of high efficient and high-density power converters, and ii) the improvement of robust and flexible modern control techniques. However, in spite of some advances, there are still open questions regarding the achievement of three of the most important issues in RE applications: reliability, efficiency, and cost. 10Renewable Energy174 1.1 DC-DC Power converters in Renewable Energy Power converters (PC) and in particular Direct Current to Direct Current (DC-DC) converters play an important role in RE since they serve as a link stage for energy conditioning (usually step-up); and they can be used in low and high power converters such as HVDC, FactS, Microgrids, Fuel cell generator, Fuel cell vehicles, Photo voltaic Panels (PVP) and hybrid generation systems*. The connection of renewable energy sources to power grids is not possible without power electronics, i.e. DC-DC converters optimize the efficiency of solar panels, rectifiers and inverters which are needed in wind generators. Automotive applications such as electric and hybrid drive trains are only possible with efficient and intelligent power electronics. X-by-wire concepts operated by power converters may lead to power saving of more than 20% (ECPE et al., 2007). Different topologies of PC are used in RE, the final decision to use one or another is based on the range of the final application, the kind of RE source, and if the system requires mandatory isolation or not (this specification varies upon country regulations). The most used topologies of DC-DC power converters are, 1. Simple buck and boost converters. 2. Interleaved and cascaded converters. 3. Synchronous converters. 4. High Frequency (HF) transformers using an isolated push-pull boost converters. 5. Full-bridge isolated converters. 6. Single-Inductor push-pull converters (SIC). 7. Double-Inductor push-pull converters (SIC). 8. Full-bridge converters. 9. Self-commutated inverters using thyristors. On the other hand, at the present time, the new directions in the design of DC-DC in RE deals with the following topics (ECPE et al., 2007) 1. System cost reduction. 2. New interconnection technologies for ultra-high power density systems and high temperature electronics. 3. Advanced thermal management; high temperature magnetics, capacitors, sensors, control ICs. 4. In lighting smart and simple dimming concepts; smart control of street lighting; high efficient. 5. Light sources (LED/OLED) and their power electronic drivers. 6. Higher level of integration e.g. for more compact energy saving lamps. 7. New topologies for photo voltaic (PV) solar converters focused on more efficient PV solar cells. 8. Load management by power electronics in distributed energy generation networks. 9. Zero-defect design and improved system reliability including fault-tolerant systems. 10. Digital power conversion and smart power management. *Here the term “hybrid” refers to the use of two different energy sources. It is in the solution of the problems related to these new tendencies in power electronics and renewable energy, that control plays a crucial role. Robust and flexible controllers are capable to guarantee efficient energy transmission and processing, are able compensate some design limitations and to reduce operation and maintenance cost. Furthermore, they can be used also for increasing reliability and safety of operation. In the past, traditional control schemes as direct duty ratio control (Mohan et al., 1989), (Rashid, 1988), (Erickson & Maksimovic, 2001), voltage and current programming control (Capel et al., 1978), (Redl & Sokal, 1985), (Middlebrook R., 1985), have been applied in the solution of some of the main operation concerns. However they are based on averaged models (Middlebrook & Cuk, 1976), (Erickson et al., 1982). Basically, an averaged model assumes high switching frequency, giving as a result continuous time models and controllers. Nevertheless, it has been shown that this procedure has some limitations in the exact estimation of the DC output voltage and harmonics components; in addition, averaged models fail to predict sub-harmonic and chaotic oscillations under some control techniques (Deane, 1990). In order to analyze this erratic behavior, it is used sometimes discrete-time maps, which relate state variables at every switching period. The main limitation of digital techniques is that obtained models are nonlinear with respect to the duty cycle (Kassakian et al., 1991); this fact makes more complicated the control design task. The highly need of flexibility and more robust controllers in areas such as motor controllers, renewable energy, and power electronics; has opened the niche to apply and develop a new family of more active control methods that are able to work in operation regions rather than in a single operation point. Control methodologies such as: fuzzy, linear quadratic, passivity based and pulse adjustment among others, are good examples of complex controllers usually designed to fulfill strict standards of performance. However, in spite of the advantages in terms of robustness and good performance, their applicability is limited by the complexity of their design and implementation. That is, in spite of the last efforts of research in the field, there is still a truly need to develop new families of controllers, due to the highly demand environments of modern power electronics applications and the advances in digital processors. Therefore, hybrid-time controller has received a lot of attention due to its practical feasibility to achieve a high performance and flexibility as well as, natural digital implementation in signal processors. There are several types of hybrid controllers that has been applied to power converters; but perhaps the most commonly found in the literature, in view of the switching nature of the converters, are the piece-wise continuous controllers. In this type of control, system commute by virtue of a discrete event, that is product of the satisfaction of a prescribed set of conditions. Due to this fact, the control is also known as switched (Liberzon, 2003), (Li et. al, 2005), (Sun & Ge, 2005) or boundary control (Ting-Ting & Chung, 2008), (Leung & Chung, 2004), (Quaicoe & Iqbal, 2008). Hybrid Control of DC-DC Power Converters 175Renewable Energy174 1.1 DC-DC Power converters in Renewable Energy Power converters (PC) and in particular Direct Current to Direct Current (DC-DC) converters play an important role in RE since they serve as a link stage for energy conditioning (usually step-up); and they can be used in low and high power converters such as HVDC, FactS, Microgrids, Fuel cell generator, Fuel cell vehicles, Photo voltaic Panels (PVP) and hybrid generation systems*. The connection of renewable energy sources to power grids is not possible without power electronics, i.e. DC-DC converters optimize the efficiency of solar panels, rectifiers and inverters which are needed in wind generators. Automotive applications such as electric and hybrid drive trains are only possible with efficient and intelligent power electronics. X-by-wire concepts operated by power converters may lead to power saving of more than 20% (ECPE et al., 2007). Different topologies of PC are used in RE, the final decision to use one or another is based on the range of the final application, the kind of RE source, and if the system requires mandatory isolation or not (this specification varies upon country regulations). The most used topologies of DC-DC power converters are, 1. Simple buck and boost converters. 2. Interleaved and cascaded converters. 3. Synchronous converters. 4. High Frequency (HF) transformers using an isolated push-pull boost converters. 5. Full-bridge isolated converters. 6. Single-Inductor push-pull converters (SIC). 7. Double-Inductor push-pull converters (SIC). 8. Full-bridge converters. 9. Self-commutated inverters using thyristors. On the other hand, at the present time, the new directions in the design of DC-DC in RE deals with the following topics (ECPE et al., 2007) 1. System cost reduction. 2. New interconnection technologies for ultra-high power density systems and high temperature electronics. 3. Advanced thermal management; high temperature magnetics, capacitors, sensors, control ICs. 4. In lighting smart and simple dimming concepts; smart control of street lighting; high efficient. 5. Light sources (LED/OLED) and their power electronic drivers. 6. Higher level of integration e.g. for more compact energy saving lamps. 7. New topologies for photo voltaic (PV) solar converters focused on more efficient PV solar cells. 8. Load management by power electronics in distributed energy generation networks. 9. Zero-defect design and improved system reliability including fault-tolerant systems. 10. Digital power conversion and smart power management. *Here the term “hybrid” refers to the use of two different energy sources. It is in the solution of the problems related to these new tendencies in power electronics and renewable energy, that control plays a crucial role. Robust and flexible controllers are capable to guarantee efficient energy transmission and processing, are able compensate some design limitations and to reduce operation and maintenance cost. Furthermore, they can be used also for increasing reliability and safety of operation. In the past, traditional control schemes as direct duty ratio control (Mohan et al., 1989), (Rashid, 1988), (Erickson & Maksimovic, 2001), voltage and current programming control (Capel et al., 1978), (Redl & Sokal, 1985), (Middlebrook R., 1985), have been applied in the solution of some of the main operation concerns. However they are based on averaged models (Middlebrook & Cuk, 1976), (Erickson et al., 1982). Basically, an averaged model assumes high switching frequency, giving as a result continuous time models and controllers. Nevertheless, it has been shown that this procedure has some limitations in the exact estimation of the DC output voltage and harmonics components; in addition, averaged models fail to predict sub-harmonic and chaotic oscillations under some control techniques (Deane, 1990). In order to analyze this erratic behavior, it is used sometimes discrete-time maps, which relate state variables at every switching period. The main limitation of digital techniques is that obtained models are nonlinear with respect to the duty cycle (Kassakian et al., 1991); this fact makes more complicated the control design task. The highly need of flexibility and more robust controllers in areas such as motor controllers, renewable energy, and power electronics; has opened the niche to apply and develop a new family of more active control methods that are able to work in operation regions rather than in a single operation point. Control methodologies such as: fuzzy, linear quadratic, passivity based and pulse adjustment among others, are good examples of complex controllers usually designed to fulfill strict standards of performance. However, in spite of the advantages in terms of robustness and good performance, their applicability is limited by the complexity of their design and implementation. That is, in spite of the last efforts of research in the field, there is still a truly need to develop new families of controllers, due to the highly demand environments of modern power electronics applications and the advances in digital processors. Therefore, hybrid-time controller has received a lot of attention due to its practical feasibility to achieve a high performance and flexibility as well as, natural digital implementation in signal processors. There are several types of hybrid controllers that has been applied to power converters; but perhaps the most commonly found in the literature, in view of the switching nature of the converters, are the piece-wise continuous controllers. In this type of control, system commute by virtue of a discrete event, that is product of the satisfaction of a prescribed set of conditions. Due to this fact, the control is also known as switched (Liberzon, 2003), (Li et. al, 2005), (Sun & Ge, 2005) or boundary control (Ting-Ting & Chung, 2008), (Leung & Chung, 2004), (Quaicoe & Iqbal, 2008). Hybrid Control of DC-DC Power Converters 175Renewable Energy176 Most switched control algorithms are easy to implement with respect to other nonlinear methodologies and they have advantage that the switching criteria can be expressed by very intuitive (design) restrictions, i.e. maximum current switching criterion, maximum ripple switching criterion, etc. Moreover, some of these controllers have high flexibility and adaptability in a large set of operating conditions, including changes on power load and voltage source or under continuous and discontinuous operation mode. Furthermore, since their design is not based on an averaged behaviour of the converter they are suitable to compensate a wide range of nonlinear phenomena. Following this trend in the development of flexible controllers, the main purpose of this chapter is to describe some of the most significant results regarding hybrid controllers used in DC-DC converters; to discuss their stability properties and their limitations, as well as their applicability and relevance. The chapter begins (Section 2) with an introduction to hybrid systems (HS). Different definitions found in the literature are discussed and special attention is paid to switched and impulsive systems. Moreover, a detailed description of the switched characteristics is given. In Section 3, a description of the tools used to analyse stability of HS is performed, emphasis is made about the class of stability result obtained from each methodology. A general description of most relevant hybrid controllers proposed in the literature is given in Section 4; it is stated that the proposed methodologies can be classified in two different kinds and particularities of each one are discussed. General aspects as robustness and implementing advantages and drawbacks are discussed. Finally some concluding remarks are given (Section 5). 2. Hybrid Systems Description The term hybrid denotes a mixed origin. In physics a hybrid electromagnetic wave is one having components of both electric and magnetic field vectors in the direction of propagation. A hybrid vehicle is one having two or more energy sources; for example electric and chemical (combustion). Regarding dynamic systems, there is not a universal definition of hybrid systems, but most authors coincide that in such systems, it exists interaction and coexistence of two kinds of dynamics: continuous and discrete (Liberzon, 2003), (Li et al., 2005) (Goebel et al., 2009). In view of this idea, most real-time controlled systems can be called hybrid; for example, an electric motor or a generator controlled by a Digital Signal Processor (DSP). In this case, the continuous dynamics is given by the machine, which by the process of sampling and data processing is controlled at discrete times. Moreover, notice that the definition above can describe interaction between finite or infinite dimensional dynamic systems with discrete events or dynamics. It can be considered as hybrid, continuous systems manipulated by piece continuous controls as supervisory control, sliding mode control and gain scheduling control for example. On the other hand, controllers as current peak control (CPC) are also a good example of hybrid controllers. Deriving a general model for hybrid systems (HS) may be quite complicated, since various types of interactions between discrete and continuous dynamics may exist. Models found in the literature vary mainly depending upon the type of HS studied and the way the author conceives such interaction. Among hybrid systems, switched and impulsive systems are probably the most studied and since the discrete interaction is well defined, general models can be defined. At this point, it is worthy noticing that some authors (Sun & Ge, 2005) consider also as impulsive and switched, classes of systems that commute between discrete dynamics. That is, no continuous dynamics exists and the overall system is constituted only by different discrete behaviours. Such systems are not hybrid in the sense discussed here, even if they are truly switching or impulsive. In this way, the terms switched and impulsive cannot be taken as synonymous of hybrid. In Impulsive systems (IS) the members of a family of continuous systems (subsystems or modes) interact between them by virtue of a discrete event or condition. Such discrete interaction changes the dynamics in such a way that the system trajectories are discontinuous. In Fig. 1 it is possible to observe the time evolution of a second order impulsive system (i.e. x = [x1,x2]T ) as well as the system trajectories in the phase plane. Observe that in a given time, the trajectories reset to a given new value. This is, the discrete interaction induces a jump on systems trajectories making the initial condition at the k+1 subsystem be different from the final condition of the k subsystem. In this way, the trajectories of an impulsive system are only piecewise continuous. Examples of impulsive systems are those describing the population of insects and other living species in which, grow rate varies impulsively (i.e. birth rate may lead to instantaneously population differences). Some chemical reactors with fast continuous dynamics behave as IS if the reactive is fed impulsively. Some financial systems behave also impulsively. Some other examples can be found in (Yang, 2001 a,b), (Li et al., 2005), (Sun & Ge, 2005). Impulsive systems are almost inexistent in DC-DC power electronics, even if perturbations are usually modelled as impulsive. The reason resides in the fact that in most cases, there exist continuous dynamics in such applications that serve as a filter (slow dynamics), making continuous every trajectory subjected to impulsive phenomena. For example, the RC tank in DC-DC converters serves as a low pass filter to reduce voltage ripple originated by the (impulsive) transition of the switch. Models of impulsive systems can be found in (Yang, 2001a) and (Yang, 2001b). 2.1 Switching Systems Since switching systems are of special interest in power electronics, this subsection will be devoted to gain insight in the most common description used to describe such systems. We begin by saying that switching systems are constituted mainly of two parts: i) The family of continuous subsystems or modes and, ii) the switching law. As in the case of hybrid impulsive systems; in switched systems, the members of a family of continuous systems (subsystems or modes) interact between them by virtue of a discrete event or condition. Hybrid Control of DC-DC Power Converters 177Renewable Energy176 Most switched control algorithms are easy to implement with respect to other nonlinear methodologies and they have advantage that the switching criteria can be expressed by very intuitive (design) restrictions, i.e. maximum current switching criterion, maximum ripple switching criterion, etc. Moreover, some of these controllers have high flexibility and adaptability in a large set of operating conditions, including changes on power load and voltage source or under continuous and discontinuous operation mode. Furthermore, since their design is not based on an averaged behaviour of the converter they are suitable to compensate a wide range of nonlinear phenomena. Following this trend in the development of flexible controllers, the main purpose of this chapter is to describe some of the most significant results regarding hybrid controllers used in DC-DC converters; to discuss their stability properties and their limitations, as well as their applicability and relevance. The chapter begins (Section 2) with an introduction to hybrid systems (HS). Different definitions found in the literature are discussed and special attention is paid to switched and impulsive systems. Moreover, a detailed description of the switched characteristics is given. In Section 3, a description of the tools used to analyse stability of HS is performed, emphasis is made about the class of stability result obtained from each methodology. A general description of most relevant hybrid controllers proposed in the literature is given in Section 4; it is stated that the proposed methodologies can be classified in two different kinds and particularities of each one are discussed. General aspects as robustness and implementing advantages and drawbacks are discussed. Finally some concluding remarks are given (Section 5). 2. Hybrid Systems Description The term hybrid denotes a mixed origin. In physics a hybrid electromagnetic wave is one having components of both electric and magnetic field vectors in the direction of propagation. A hybrid vehicle is one having two or more energy sources; for example electric and chemical (combustion). Regarding dynamic systems, there is not a universal definition of hybrid systems, but most authors coincide that in such systems, it exists interaction and coexistence of two kinds of dynamics: continuous and discrete (Liberzon, 2003), (Li et al., 2005) (Goebel et al., 2009). In view of this idea, most real-time controlled systems can be called hybrid; for example, an electric motor or a generator controlled by a Digital Signal Processor (DSP). In this case, the continuous dynamics is given by the machine, which by the process of sampling and data processing is controlled at discrete times. Moreover, notice that the definition above can describe interaction between finite or infinite dimensional dynamic systems with discrete events or dynamics. It can be considered as hybrid, continuous systems manipulated by piece continuous controls as supervisory control, sliding mode control and gain scheduling control for example. On the other hand, controllers as current peak control (CPC) are also a good example of hybrid controllers. Deriving a general model for hybrid systems (HS) may be quite complicated, since various types of interactions between discrete and continuous dynamics may exist. Models found in the literature vary mainly depending upon the type of HS studied and the way the author conceives such interaction. Among hybrid systems, switched and impulsive systems are probably the most studied and since the discrete interaction is well defined, general models can be defined. At this point, it is worthy noticing that some authors (Sun & Ge, 2005) consider also as impulsive and switched, classes of systems that commute between discrete dynamics. That is, no continuous dynamics exists and the overall system is constituted only by different discrete behaviours. Such systems are not hybrid in the sense discussed here, even if they are truly switching or impulsive. In this way, the terms switched and impulsive cannot be taken as synonymous of hybrid. In Impulsive systems (IS) the members of a family of continuous systems (subsystems or modes) interact between them by virtue of a discrete event or condition. Such discrete interaction changes the dynamics in such a way that the system trajectories are discontinuous. In Fig. 1 it is possible to observe the time evolution of a second order impulsive system (i.e. x = [x1,x2]T ) as well as the system trajectories in the phase plane. Observe that in a given time, the trajectories reset to a given new value. This is, the discrete interaction induces a jump on systems trajectories making the initial condition at the k+1 subsystem be different from the final condition of the k subsystem. In this way, the trajectories of an impulsive system are only piecewise continuous. Examples of impulsive systems are those describing the population of insects and other living species in which, grow rate varies impulsively (i.e. birth rate may lead to instantaneously population differences). Some chemical reactors with fast continuous dynamics behave as IS if the reactive is fed impulsively. Some financial systems behave also impulsively. Some other examples can be found in (Yang, 2001 a,b), (Li et al., 2005), (Sun & Ge, 2005). Impulsive systems are almost inexistent in DC-DC power electronics, even if perturbations are usually modelled as impulsive. The reason resides in the fact that in most cases, there exist continuous dynamics in such applications that serve as a filter (slow dynamics), making continuous every trajectory subjected to impulsive phenomena. For example, the RC tank in DC-DC converters serves as a low pass filter to reduce voltage ripple originated by the (impulsive) transition of the switch. Models of impulsive systems can be found in (Yang, 2001a) and (Yang, 2001b). 2.1 Switching Systems Since switching systems are of special interest in power electronics, this subsection will be devoted to gain insight in the most common description used to describe such systems. We begin by saying that switching systems are constituted mainly of two parts: i) The family of continuous subsystems or modes and, ii) the switching law. As in the case of hybrid impulsive systems; in switched systems, the members of a family of continuous systems (subsystems or modes) interact between them by virtue of a discrete event or condition. Hybrid Control of DC-DC Power Converters 177Renewable Energy178 Such discrete interaction changes the dynamics in such a way that the system trajectories remain continuous (see Fig 2). Good examples of switched systems are the piecewise linear systems as the following:  pmuBxAx mmm ,...1 (1) where x n is the state variable vector, ui is the input vector (possible controlled) and matrices Am and Bm have suitable dimensions. The family of continuous time systems (1) is subjected to a discrete time event m that defines the current election of the active subsystem i, that is, m is an index that stands for the switch between stages. Such an index can be described by algebraic equations, difference equations or by a set of restrictions or conditions depending possible on systems states and/ or time. The expression that defines m at every instant is called switching law. In most cases, such an expression is a single valued function at every time (i.e. uniquely defined at every time). Fig. 1. Typical evolution of a impulsive system (left) time evolution, (right) phase portrait. The switching law can be state dependent and/or time dependent (Liberzon, 2003). Usually, the switching law is given by a set of prescribed conditions leading to a division of the state space. As time evolves, the switching law generates a sequence of subsystems active for a certain time, that is       kk iii ,,,, 1100   (2) where k stands for the residence time in mode or subsystem ik for ik = 1,2…p and k =1,2,3… Such a sequence is called switching sequence and it is used some times instead of the switching law to study the stability of the overall hybrid system (Xu, 2004), (Xu et al., 2007). 0 2 4 6 8 10 12 14 16 18 20-40-30-20-10010203040txx1(t)x2(t)-40 -30 -20 -10 0 10 20 30 40-20-15-10-505101520x1x 2 Fig. 2. Typical evolution of a switched system (left) time evolution, (right) phase portrait. In general, switching systems can be described as:  piutxfx i ,...1),,(  (3) where x n is the state variable vector and u is the input vector (possible controlled). On the other hand, the switching law is sometimes described as a piece-wise continuous function that in a general case depends on system states, time and the input vector as follows: )),(),(),(()( ttulitxgli  (4) where i(l+) = i(l+1). Eq. (4) is not the only way of representing a switching law; a combination of a switching criterion and mode assignment criterion is also used. The switching criterion or switching conditions are usually formulated only upon a part of the state space (output) in the form of hyper-surfaces that are called switching surfaces. It is required that every region in the entire state space has assigned an active subsystem in order to have a well defined system (Liberzon, 2003). Once the system hits a pre-defined surface (Eq. (5)), the mode assignment criterion (Eq. (6)), decides based on the function , which system is the next to be active in view of the current subsystem (Li et al., 2005). In such a case, the switching law can be described as follows: 0)),(),((  ttitxSS (5) ))(()( lili  (6) Notice that in general, a finite set of surfaces instead a single surface may be used to describe the switching criterion. In this case, the state space is divided in a finite number of regions and the possible combinations of active subsystems can increase dramatically. It is worthy to notice that hybrid switched control can be seen as a generalization of sliding mode control. This last is a switching control that uses a single switching surface along with a mode assignment criterion that chooses only between two options (Barnejee & Vergese, 2001). Hybrid Control of DC-DC Power Converters 179x x 2Renewable Energy178 Such discrete interaction changes the dynamics in such a way that the system trajectories remain continuous (see Fig 2). Good examples of switched systems are the piecewise linear systems as the following:  pmuBxAx mmm ,...1 (1) where x n is the state variable vector, ui is the input vector (possible controlled) and matrices Am and Bm have suitable dimensions. The family of continuous time systems (1) is subjected to a discrete time event m that defines the current election of the active subsystem i, that is, m is an index that stands for the switch between stages. Such an index can be described by algebraic equations, difference equations or by a set of restrictions or conditions depending possible on systems states and/ or time. The expression that defines m at every instant is called switching law. In most cases, such an expression is a single valued function at every time (i.e. uniquely defined at every time). Fig. 1. Typical evolution of a impulsive system (left) time evolution, (right) phase portrait. The switching law can be state dependent and/or time dependent (Liberzon, 2003). Usually, the switching law is given by a set of prescribed conditions leading to a division of the state space. As time evolves, the switching law generates a sequence of subsystems active for a certain time, that is       kk iii ,,,, 1100   (2) where k stands for the residence time in mode or subsystem ik for ik = 1,2…p and k =1,2,3… Such a sequence is called switching sequence and it is used some times instead of the switching law to study the stability of the overall hybrid system (Xu, 2004), (Xu et al., 2007). 0 2 4 6 8 10 12 14 16 18 20-40-30-20-10010203040txx1(t)x2(t)-40 -30 -20 -10 0 10 20 30 40-20-15-10-505101520x1x 2 Fig. 2. Typical evolution of a switched system (left) time evolution, (right) phase portrait. In general, switching systems can be described as:  piutxfx i ,...1),,(  (3) where x n is the state variable vector and u is the input vector (possible controlled). On the other hand, the switching law is sometimes described as a piece-wise continuous function that in a general case depends on system states, time and the input vector as follows: )),(),(),(()( ttulitxgli  (4) where i(l+) = i(l+1). Eq. (4) is not the only way of representing a switching law; a combination of a switching criterion and mode assignment criterion is also used. The switching criterion or switching conditions are usually formulated only upon a part of the state space (output) in the form of hyper-surfaces that are called switching surfaces. It is required that every region in the entire state space has assigned an active subsystem in order to have a well defined system (Liberzon, 2003). Once the system hits a pre-defined surface (Eq. (5)), the mode assignment criterion (Eq. (6)), decides based on the function , which system is the next to be active in view of the current subsystem (Li et al., 2005). In such a case, the switching law can be described as follows: 0)),(),((  ttitxSS (5) ))(()( lili  (6) Notice that in general, a finite set of surfaces instead a single surface may be used to describe the switching criterion. In this case, the state space is divided in a finite number of regions and the possible combinations of active subsystems can increase dramatically. It is worthy to notice that hybrid switched control can be seen as a generalization of sliding mode control. This last is a switching control that uses a single switching surface along with a mode assignment criterion that chooses only between two options (Barnejee & Vergese, 2001). Hybrid Control of DC-DC Power Converters 179x x 2Renewable Energy180 Notice that in Eqs. (3)-(4), there exists the possibility of using different control inputs in a switching system. That is, a control problem in a general switched system can be formulated as one of stabilization or tracking, using three kinds of control inputs: The switching criterion, the mode assignment criterion and the continuous control input u. To be precise, using hybrid control one can increase the number of degrees of freedom that can be used to accomplish a control objective. There exist in the control literature, various authors leading with these control problems (Lin, 2008), (Lazar, et al., 2006), (Tsinias, (2007), (Zhai & Michel, 2002), (Zhai & Michel, 2003), (Rodrigues & How, 2003). However much of this literature cannot be applied directly to power electronics due that the input u cannot be manipulated (consider converter description as in (3), (4)). In other words, the vector fields among which the converter is evolving between switches cannot be manipulated; hence, properties as stability and equilibrium points can not be modified. In view to this fact, a relevant problem in switching control applied to power electronics is how to choose the switching surfaces (possible moving surfaces) in order to stabilize the overall system. Observe that, given a mode assignment criterion, the manipulation of surfaces can be seen as equivalent to choose a set of switching times. Since in DC-DC Power electronics, the control actions are applied only at some instants (not necessarily synchronous), sometimes these systems are called partially open-loop systems. Another option of describing switched systems can be performed using Hybrid Automata. A hybrid automaton (HA) is a mathematical model, borrowed from computation theory, used to describe processes consisting in discrete state transition and continuous evolution. In this way, HA can be seen as a finite state machine that uses differential equations; in which the model is composed of a finite number of states, transitions between states and actions (differential equations) (Henzinger, 1995), (Henzinger, 1996), (Rajeev et al., 1996). A hybrid automaton H is composed by a finite set of variables (X = {x1, x2…xn}), that describe the system during continuous evolution (X = {x1, x2…xn}) and during discrete transitions (X´ = {x’1, x’2…x’n}). The system dynamics is represented by a multi-directed graph called a control graph (V,E). The vertices V are called control modes. The edges E are called control switches or guards. By virtue of a jump or switch condition it is assigned to each control switch e  E a given consequence which is detonated by an event. In mathematics literature, the description of switched system by means of differential inclusions is common; however, in power electronics literature is rarely found. The interested reader is referred to (Filipov, 1998) for detailed description of this modelling tool. 2.2 Examples in DC_DC Power converters In order to illustrate the modelling techniques mentioned above, the case of a conventional boost converter is used as a benchmark. In this way, the piece-wise continuous equations describing the converter dynamics are given by Eq (1) where x = [iL,vc]2 is the state vector constituted by the inductor current and the output voltage, u2 is the input vector that account for the translation due to the input voltage source. If the converter is working under continuous conduction mode (CCM) the number of modes or subsystems is p = 2 with 0000111LERCBandA for m=1 (7) 0021112LERCCL BandA for m=2 where E is the voltage input, L is the inductance, C the capacitance and R the load resistance. If the converter is working under discontinuous conduction mode (DCM) the number of modes or subsystems is p = 3 with m = 1 and m = 2 given as in Eq. (7) and 00000313 BandARC for m=3 (8) The description of the switching law will depend if the converter is open or closed loop and on the switching regime used. If a driven signal of frequency f = 1/T is used, the switching law can be described as follows for a converter operating in CCM: TTtTifTTtifti),mod(2),mod(01)( (9) where the function mod(a,b) stands for the remainder of the division of a/b, and  is the duty cycle. For the converter in DCM the switching law is given by TTtandTTtTifTTtandTTtTifTTtifti),mod(),mod(3),mod(),mod(2),mod(01)(22 (10) where T-T2 stands for the time length that the inductor has zero current. On the other hand, the description of the converter in both CCM and DCM using hybrid automaton can be observed in the control graph shown in Fig. 3, which is a reproduction of Figure 3 in (Sreekumar & Agarwal, 2008), where the inductor currents rising phase is defined with 1†, the falling phase with 2 and zero phase with 3. The transition between stages () is defined by the guards in Table 1. In Fig. 3, q are the discrete states, x are the continuous states, I1,2,3 are invariant sets and G are the guards. † m = 1 in Eq. (7) Hybrid Control of DC-DC Power Converters 181Renewable Energy180 Notice that in Eqs. (3)-(4), there exists the possibility of using different control inputs in a switching system. That is, a control problem in a general switched system can be formulated as one of stabilization or tracking, using three kinds of control inputs: The switching criterion, the mode assignment criterion and the continuous control input u. To be precise, using hybrid control one can increase the number of degrees of freedom that can be used to accomplish a control objective. There exist in the control literature, various authors leading with these control problems (Lin, 2008), (Lazar, et al., 2006), (Tsinias, (2007), (Zhai & Michel, 2002), (Zhai & Michel, 2003), (Rodrigues & How, 2003). However much of this literature cannot be applied directly to power electronics due that the input u cannot be manipulated (consider converter description as in (3), (4)). In other words, the vector fields among which the converter is evolving between switches cannot be manipulated; hence, properties as stability and equilibrium points can not be modified. In view to this fact, a relevant problem in switching control applied to power electronics is how to choose the switching surfaces (possible moving surfaces) in order to stabilize the overall system. Observe that, given a mode assignment criterion, the manipulation of surfaces can be seen as equivalent to choose a set of switching times. Since in DC-DC Power electronics, the control actions are applied only at some instants (not necessarily synchronous), sometimes these systems are called partially open-loop systems. Another option of describing switched systems can be performed using Hybrid Automata. A hybrid automaton (HA) is a mathematical model, borrowed from computation theory, used to describe processes consisting in discrete state transition and continuous evolution. In this way, HA can be seen as a finite state machine that uses differential equations; in which the model is composed of a finite number of states, transitions between states and actions (differential equations) (Henzinger, 1995), (Henzinger, 1996), (Rajeev et al., 1996). A hybrid automaton H is composed by a finite set of variables (X = {x1, x2…xn}), that describe the system during continuous evolution (X = {x1, x2…xn}) and during discrete transitions (X´ = {x’1, x’2…x’n}). The system dynamics is represented by a multi-directed graph called a control graph (V,E). The vertices V are called control modes. The edges E are called control switches or guards. By virtue of a jump or switch condition it is assigned to each control switch e  E a given consequence which is detonated by an event. In mathematics literature, the description of switched system by means of differential inclusions is common; however, in power electronics literature is rarely found. The interested reader is referred to (Filipov, 1998) for detailed description of this modelling tool. 2.2 Examples in DC_DC Power converters In order to illustrate the modelling techniques mentioned above, the case of a conventional boost converter is used as a benchmark. In this way, the piece-wise continuous equations describing the converter dynamics are given by Eq (1) where x = [iL,vc]2 is the state vector constituted by the inductor current and the output voltage, u2 is the input vector that account for the translation due to the input voltage source. If the converter is working under continuous conduction mode (CCM) the number of modes or subsystems is p = 2 with 0000111LERCBandA for m=1 (7) 0021112LERCCL BandA for m=2 where E is the voltage input, L is the inductance, C the capacitance and R the load resistance. If the converter is working under discontinuous conduction mode (DCM) the number of modes or subsystems is p = 3 with m = 1 and m = 2 given as in Eq. (7) and 00000313 BandARC for m=3 (8) The description of the switching law will depend if the converter is open or closed loop and on the switching regime used. If a driven signal of frequency f = 1/T is used, the switching law can be described as follows for a converter operating in CCM: TTtTifTTtifti),mod(2),mod(01)( (9) where the function mod(a,b) stands for the remainder of the division of a/b, and  is the duty cycle. For the converter in DCM the switching law is given by TTtandTTtTifTTtandTTtTifTTtifti),mod(),mod(3),mod(),mod(2),mod(01)(22 (10) where T-T2 stands for the time length that the inductor has zero current. On the other hand, the description of the converter in both CCM and DCM using hybrid automaton can be observed in the control graph shown in Fig. 3, which is a reproduction of Figure 3 in (Sreekumar & Agarwal, 2008), where the inductor currents rising phase is defined with 1†, the falling phase with 2 and zero phase with 3. The transition between stages () is defined by the guards in Table 1. In Fig. 3, q are the discrete states, x are the continuous states, I1,2,3 are invariant sets and G are the guards. † m = 1 in Eq. (7) Hybrid Control of DC-DC Power Converters 181Renewable Energy182 Fig. 3. Hybrid description of a conventional boost converter under CCM and DCM (Sreekumar & Agarwal, 2008). where: G12 0T G21 T T Table 1. Guard description of Fig. 3. Notice that under CCM the active guards are G12 and G21. Moreover, under DCM the switching can only be performed unidirectionally; that is G12 → G23 → G31 → G12 →… (Senesky et al., 2003). 2.3 Control Problems in DC-DC Power Converters. Hybrid Control Solution This subsection is devoted to describe the control problems in DC-DC power converters that are best suitable to be addressed with hybrid control, as well as describe some advantages of this tool. From conceiving the system as piecewise continuous one or by using a combination of classical averaged models with discrete transitions or discrete controllers is that hybrid control arises in power electronics. That is, the system under study has open and/or closed loop hybrid description and, as stated before, its mathematical description can be very vast due to the variety of the possible interactions. In order to describe the control problems in DC-DC power converters, it is worthy noticing that the main performance concerns in these applications are: a. Voltage regulation. Voltage regulation of a constant desired reference is the primary objective of control and a basic performance requirement that has to be fulfilled in every real life application. b. Current limitation. For safety reasons and to endure converter components life, it is required a current limitation in the circuit. c. Maximum current and voltage ripple. Between 10-20% of the reference current(s) signal(s) and about 1-2% of the desired voltage. Furthermore, it is highly desirable that the control is designed to fulfil the following requirements:  Adaptability or Applicability to CCM or DCM regimes. Flexibility to operate in both modes, mainly due to possible perturbations in the parameters and to face near- to- cero load resistance conditions.  Robustness against model uncertainty and disturbances. For example, disturbances in load resistance, input voltage and voltage reference. That is, it is desirable that the converter operates in regions of the state space, rather than in a small neighborhood about the nominal point.  Output feedback. It is desirable that the control uses only a subset of the state vector for feedback purposes (output feedback). This is of particular interest in high order converters, where measuring all signals becomes prohibitive due to the high costs.  Fixed Switched Frequency. Among the main reasons to operate at fixed frequency conditions are: i) the possibility that harmonic content arise in varying frequency operation, which in turn may induce a poor steady-state performance and ii) It is easiest to implement fixed frequency controllers. An advantage of hybrid control and in particular, of switched control is that some performance requirements can be formulated explicitly by defining suitable switching surfaces or control guards. For example, effective current limitation can be performed with a time varying current-dependent switching surface i.e. peak current control with stabilizing ramp; voltage or current ripple can be limited for a suitable hysteresis-like controller , which can be seen as a switching control with two voltage and/or current dependent switching surfaces (Cervantes et al., 2009a), (Perez-Pinal & Cervantes, 2009a). Furthermore, due to the fact that the converters can be seen as switched systems, the switching criterion and the mode assignment criterion can be manipulated independently to obtain multiple control objective, for example, current limitation at the same time that voltage regulation (Austria-Gonzalez et al., 2008) (Perez-Pinal & Cervantes, 2009b), or maximum voltage and current rippling as well as voltage regulation (Cervantes et al., 2009b). On the other hand, hybrid controllers represent an option to solve optimal operation problems. For example in (Almer et al., 2007) a predictive control with load estimation is proposed, in which a performance index is minimized to obtain a prescribed performance (see also, Geyer et al., 2008). In (Chen et al., 2006) a control strategy is proposed that is able to compute optimal switching instances that delivered over one switching period allow to a series parallel resonant converter to achieve its required steady state in minimum time. Hybrid controllers can also be designed to solve voltage regulation problems even if the converter changes from CCM to DCM regime and vice versa (Sreekumar &, Agarwal, 2006), (Sreekumar &, Agarwal, 2007), (Sreekumar &, Agarwal, 2008), (Senesky et al., 2003). Furthermore, many of the hybrid control strategies proposed can be implemented using fixed frequency switching (Geyer et al., 2008). Hybrid Control of DC-DC Power Converters 183Renewable Energy182 Fig. 3. Hybrid description of a conventional boost converter under CCM and DCM (Sreekumar & Agarwal, 2008). where: G12 0T G21 T T Table 1. Guard description of Fig. 3. Notice that under CCM the active guards are G12 and G21. Moreover, under DCM the switching can only be performed unidirectionally; that is G12 → G23 → G31 → G12 →… (Senesky et al., 2003). 2.3 Control Problems in DC-DC Power Converters. Hybrid Control Solution This subsection is devoted to describe the control problems in DC-DC power converters that are best suitable to be addressed with hybrid control, as well as describe some advantages of this tool. From conceiving the system as piecewise continuous one or by using a combination of classical averaged models with discrete transitions or discrete controllers is that hybrid control arises in power electronics. That is, the system under study has open and/or closed loop hybrid description and, as stated before, its mathematical description can be very vast due to the variety of the possible interactions. In order to describe the control problems in DC-DC power converters, it is worthy noticing that the main performance concerns in these applications are: a. Voltage regulation. Voltage regulation of a constant desired reference is the primary objective of control and a basic performance requirement that has to be fulfilled in every real life application. b. Current limitation. For safety reasons and to endure converter components life, it is required a current limitation in the circuit. c. Maximum current and voltage ripple. Between 10-20% of the reference current(s) signal(s) and about 1-2% of the desired voltage. Furthermore, it is highly desirable that the control is designed to fulfil the following requirements:  Adaptability or Applicability to CCM or DCM regimes. Flexibility to operate in both modes, mainly due to possible perturbations in the parameters and to face near- to- cero load resistance conditions.  Robustness against model uncertainty and disturbances. For example, disturbances in load resistance, input voltage and voltage reference. That is, it is desirable that the converter operates in regions of the state space, rather than in a small neighborhood about the nominal point.  Output feedback. It is desirable that the control uses only a subset of the state vector for feedback purposes (output feedback). This is of particular interest in high order converters, where measuring all signals becomes prohibitive due to the high costs.  Fixed Switched Frequency. Among the main reasons to operate at fixed frequency conditions are: i) the possibility that harmonic content arise in varying frequency operation, which in turn may induce a poor steady-state performance and ii) It is easiest to implement fixed frequency controllers. An advantage of hybrid control and in particular, of switched control is that some performance requirements can be formulated explicitly by defining suitable switching surfaces or control guards. For example, effective current limitation can be performed with a time varying current-dependent switching surface i.e. peak current control with stabilizing ramp; voltage or current ripple can be limited for a suitable hysteresis-like controller , which can be seen as a switching control with two voltage and/or current dependent switching surfaces (Cervantes et al., 2009a), (Perez-Pinal & Cervantes, 2009a). Furthermore, due to the fact that the converters can be seen as switched systems, the switching criterion and the mode assignment criterion can be manipulated independently to obtain multiple control objective, for example, current limitation at the same time that voltage regulation (Austria-Gonzalez et al., 2008) (Perez-Pinal & Cervantes, 2009b), or maximum voltage and current rippling as well as voltage regulation (Cervantes et al., 2009b). On the other hand, hybrid controllers represent an option to solve optimal operation problems. For example in (Almer et al., 2007) a predictive control with load estimation is proposed, in which a performance index is minimized to obtain a prescribed performance (see also, Geyer et al., 2008). In (Chen et al., 2006) a control strategy is proposed that is able to compute optimal switching instances that delivered over one switching period allow to a series parallel resonant converter to achieve its required steady state in minimum time. Hybrid controllers can also be designed to solve voltage regulation problems even if the converter changes from CCM to DCM regime and vice versa (Sreekumar &, Agarwal, 2006), (Sreekumar &, Agarwal, 2007), (Sreekumar &, Agarwal, 2008), (Senesky et al., 2003). Furthermore, many of the hybrid control strategies proposed can be implemented using fixed frequency switching (Geyer et al., 2008). Hybrid Control of DC-DC Power Converters 183Renewable Energy184 Moreover, an additional advantage of describing the converter dynamics using a piece-wise model allows a confident description of the system for the whole operating regime. This fact opens the possibility of designing control laws valid in every operation point. 3. Some Stability Issues A main concern when dealing with converter control is the closed-loop stability of the system. Regarding stability of hybrid systems, it is possible to find various approximations in the literature (Aeyels & Peuteman, 1998), (Branicky, 1998), (Filipov, 1998), (Liberzon, 2003) among many others. However, a great percentage of those results cannot be used directly in power electronics due to the fact that the HS considered in such works, do not fit in a possible description of a converter. That is, it is considered one or more of the following assumptions,  Vector fields of the continuous part of the HS can be modified.  In switched systems, o Subsystems have a common equilibrium point o All subsystems have equilibrium points o Subsystems have a given stability property The reasons of the applicability limitation are clear when one looks into the converter structure. Consider for example the boost converter in Eqs. (7)-(8) subsystems 1 and 3 do not have any equilibrium point. Theoretically, the current is able to grow unlimited, while subsystem 2 has an stable equilibrium point in x = (E/R ,E)T. Therefore, common equilibrium point approaches cannot be applied. Furthermore, as discussed before, the vector fields among which, the converter is evolving (subsystems 1-3) cannot be manipulated, since an extra control input u in Eq. (3) is not available. That is, neither equilibrium points nor stability properties of the subsystems can be modified. In view of this fact and as stated before, admissible control inputs are the residence time in each subsystem and mode assignment criterion. 3.1 Asymptotic Stability vs. Practical Stability It is possible to find various definitions of stability in the literature; each one describing different behaviour of the system. Boundedness, input-output stability, output-state stability, asymptotic stability, exponential stability, practical stability and absolute stability are examples of these concepts. Although the objective of this subsection is not giving an exhaust description of such notions (the interested reader is referred to (Sontag, 1998) and (Khalil, 2002), it is relevant to notice that the inexistence or existence of equilibrium points is crucial to determine the type of stability that can be achieved by the system. For example, asymptotic stability, exponential stability and absolute stability can only be achieved by systems with equilibrium points. In fact, these concepts are related to the ability of system trajectories to approach and stay at the equilibrium point (EP). In this way, it is said that trajectories are attracted by the EP and the set of all initial conditions such that this happens is called attractive domain or region of attraction. Having a large region of attraction is of great interest to controller designers mainly because it guarantees the existence of a set about the equilibrium point such that system trajectories are confined. This fact gives to the system, natural robustness against some perturbations. Although it would be desirable to obtain asymptotic (or exponential) stability in power converts, in most cases this cannot be achieved using a hybrid description. The reason resides in the fact that in most cases, a common equilibrium point can not be found, e.g. the hybrid system (3)-(4). At this point, it is interesting to notice that when an average converter model is used along linear or nonlinear continuous controllers, a common average equilibrium point can be obtained. From the continuous point of view, we may be able to achieve both asymptotic and exponential stability to this “fictitious” equilibrium point; however in the converter, attractivity to a neighbourhood of this averaged equilibrium can only be obtained (practical stability). Due to this fact, many of the results in switched and hybrid systems deal with the concept of practical stability, rather than with asymptotic stability. Since these two concepts will appear frequently in this chapter, the definitions are stated below: Definition 1. (Khalil, 2002). The equilibrium point x = 0 is of the system ),( txFx  is  stable if, for each 0 , there is 0),( 0  t such that 0,)()( 00  tttxtx   asymptotically stable if it stable and there is a positive constant )( 0tcc  such that 0)( tx as t for all ctx )( 0 Definition 2. (Xu et al., 2007). Given 0 , system (3) with u=0,  is -practically uniformly stable (-PUS) about the origin under (4), if there is a )(  >0, independent of 0t such that 0),0()(),0()( 00  ttBtxBtx   is -practically uniformly asymptotically stable (-PUAS) about the origin under (4), if is -practically uniformly stable and there is a 0)(  cTT independent of 0t such that TttBtxDcBtx  00 ),,0()(),0()(  Notice that the stability definition 1 differs from the concept of -PUS in the fact that in Definition 1 the condition over  is assumed for all 0 , while in Definition 2, is enough the existence of some value (usually design-fixed). Furthermore, notice that Definition 1 requires the existence of an equilibrium point and the stability property is attributed to it. In contrast, in Definition 2, is the system the one which is stable and the result basically establishes confinement of system trajectories about the origin and in -PUAS, also attractiveness. On the other hand, when dealing with switched control, the concept of practical stability is also related with the controlled invariant one (Senesky et al., 2003). Notice that Definition 1 concerns about continuous nonlinear, time varying systems. If the hybrid system can be described as in this definition, necessary and/or sufficient conditions can be found to deduce stability (Sontag, 1998), (Khalil, 2002) . On the other hand, conditions Hybrid Control of DC-DC Power Converters 185Renewable Energy184 Moreover, an additional advantage of describing the converter dynamics using a piece-wise model allows a confident description of the system for the whole operating regime. This fact opens the possibility of designing control laws valid in every operation point. 3. Some Stability Issues A main concern when dealing with converter control is the closed-loop stability of the system. Regarding stability of hybrid systems, it is possible to find various approximations in the literature (Aeyels & Peuteman, 1998), (Branicky, 1998), (Filipov, 1998), (Liberzon, 2003) among many others. However, a great percentage of those results cannot be used directly in power electronics due to the fact that the HS considered in such works, do not fit in a possible description of a converter. That is, it is considered one or more of the following assumptions,  Vector fields of the continuous part of the HS can be modified.  In switched systems, o Subsystems have a common equilibrium point o All subsystems have equilibrium points o Subsystems have a given stability property The reasons of the applicability limitation are clear when one looks into the converter structure. Consider for example the boost converter in Eqs. (7)-(8) subsystems 1 and 3 do not have any equilibrium point. Theoretically, the current is able to grow unlimited, while subsystem 2 has an stable equilibrium point in x = (E/R ,E)T. Therefore, common equilibrium point approaches cannot be applied. Furthermore, as discussed before, the vector fields among which, the converter is evolving (subsystems 1-3) cannot be manipulated, since an extra control input u in Eq. (3) is not available. That is, neither equilibrium points nor stability properties of the subsystems can be modified. In view of this fact and as stated before, admissible control inputs are the residence time in each subsystem and mode assignment criterion. 3.1 Asymptotic Stability vs. Practical Stability It is possible to find various definitions of stability in the literature; each one describing different behaviour of the system. Boundedness, input-output stability, output-state stability, asymptotic stability, exponential stability, practical stability and absolute stability are examples of these concepts. Although the objective of this subsection is not giving an exhaust description of such notions (the interested reader is referred to (Sontag, 1998) and (Khalil, 2002), it is relevant to notice that the inexistence or existence of equilibrium points is crucial to determine the type of stability that can be achieved by the system. For example, asymptotic stability, exponential stability and absolute stability can only be achieved by systems with equilibrium points. In fact, these concepts are related to the ability of system trajectories to approach and stay at the equilibrium point (EP). In this way, it is said that trajectories are attracted by the EP and the set of all initial conditions such that this happens is called attractive domain or region of attraction. Having a large region of attraction is of great interest to controller designers mainly because it guarantees the existence of a set about the equilibrium point such that system trajectories are confined. This fact gives to the system, natural robustness against some perturbations. Although it would be desirable to obtain asymptotic (or exponential) stability in power converts, in most cases this cannot be achieved using a hybrid description. The reason resides in the fact that in most cases, a common equilibrium point can not be found, e.g. the hybrid system (3)-(4). At this point, it is interesting to notice that when an average converter model is used along linear or nonlinear continuous controllers, a common average equilibrium point can be obtained. From the continuous point of view, we may be able to achieve both asymptotic and exponential stability to this “fictitious” equilibrium point; however in the converter, attractivity to a neighbourhood of this averaged equilibrium can only be obtained (practical stability). Due to this fact, many of the results in switched and hybrid systems deal with the concept of practical stability, rather than with asymptotic stability. Since these two concepts will appear frequently in this chapter, the definitions are stated below: Definition 1. (Khalil, 2002). The equilibrium point x = 0 is of the system ),( txFx  is  stable if, for each 0 , there is 0),( 0  t such that 0,)()( 00  tttxtx   asymptotically stable if it stable and there is a positive constant )( 0tcc  such that 0)( tx as t for all ctx )( 0 Definition 2. (Xu et al., 2007). Given 0 , system (3) with u=0,  is -practically uniformly stable (-PUS) about the origin under (4), if there is a )(  >0, independent of 0t such that 0),0()(),0()( 00  ttBtxBtx   is -practically uniformly asymptotically stable (-PUAS) about the origin under (4), if is -practically uniformly stable and there is a 0)(  cTT independent of 0t such that TttBtxDcBtx  00 ),,0()(),0()(  Notice that the stability definition 1 differs from the concept of -PUS in the fact that in Definition 1 the condition over  is assumed for all 0 , while in Definition 2, is enough the existence of some value (usually design-fixed). Furthermore, notice that Definition 1 requires the existence of an equilibrium point and the stability property is attributed to it. In contrast, in Definition 2, is the system the one which is stable and the result basically establishes confinement of system trajectories about the origin and in -PUAS, also attractiveness. On the other hand, when dealing with switched control, the concept of practical stability is also related with the controlled invariant one (Senesky et al., 2003). Notice that Definition 1 concerns about continuous nonlinear, time varying systems. If the hybrid system can be described as in this definition, necessary and/or sufficient conditions can be found to deduce stability (Sontag, 1998), (Khalil, 2002) . On the other hand, conditions Hybrid Control of DC-DC Power Converters 185Renewable Energy186 that lead us to conclude asymptotic stability of a general HS may be hard to find. However, for switched systems such a task is easier in some cases.‡ In the following subsection we will describe some contributions in the area of hybrid control applied to DC-DC power converter. The discussion about their stability properties will be explained in view of the concepts and ideas stated here. 4. Types of HC in Power Electronics The types of HC found in the literature can be classified in two groups mainly; a first one uses either a switched description or a switching control action in the converter; in these papers, the control objective is to find a switching law such that voltage is tracked. A second group, which description of the converter may be continuous or piece wise continuous, in these strategies the control objective is aimed to solve optimally (time or using and index performance) the voltage regulation problem. We will refer to the first group as Hybrid Switched Control (HSC) and to the second as, Hybrid Optimal Control (HOBC). 4.1 Hybrid Switched Control In this category we can find variable structure controllers such as: hysteresis, sliding and current peak mode controllers among the most common. Sliding mode controllers (SMC) are usually designed based on an averaged continuous model of the converter. In general, the converter is described by non-linear differential equations evolving in a fictitious averaged state space. Under this situation the control input is also continuous but restricted to  1,0u . Once a continuous model is obtained, the next step is to define a suitable surface that resembles the desired steady state, usually such surface is a plane of the form: niii xaS1~ (11) where x~ =x-xref. and xref the desired state. However, the sliding surface can be also non-linear (Sira-Ramirez & Ríos-Bolivar, 1994), (Tokat et al., 2002). Sliding mode control action can be divided in two operation modes: the reaching mode and the sliding mode. The reaching mode is related with the operation phase in which from an arbitrary initial condition the trajectories goes toward the surface S. The control input of SMC can be seen also as divided in two, ueq that acts mostly on the reaching mode and ud for the sliding. The input ueq is usually continuous and parameter dependent, while ud is a discontinuous input usually depending on the sign of the desired surface therefore, is parameter insensitive. There exist various papers in the literature dealing with the robust design of controllers for the reaching mode. For example in (Sira-Ramirez & Ríos-Bolivar, 1994) a passivity based controller is proposed to enhance the robustness properties of SMC. The controller make use of the passive relation between input voltage and inductor current of the converter, along with ‡There exist trivial cases regarding stability of switched systems that are not of our interest. For example, if there is a stable mode in the subsystems family, we can guarantee stability of the switched system just by staying indefinitely on it, among others cases. Lyapunov based arguments to show that the desired tracking dynamics approaches the origin asymptotically (i.e. the equilibrium point is asymptotically stable). The use of non-linear and time-varying sliding surfaces has also been proposed to enhance the robustness of SMC (Sira-Ramirez & Ríos-Bolivar, 1994), (Sira-Ramirez et al., 1995), (Tokat et al., 2002). On the other hand, sensorless SMC has also been proposed, as well as some techniques for fixed frequency operation, the interested reader is refer to (Tan et al., 2008) for a comprehensive review of sliding mode control in power electronics. Fig. 4. Types of hybrid controllers used in DC-DC power converters. Hysteresis control can be interpreted as a switched control with two output-dependent switching surfaces: turn-on and turn-off surfaces (Babaa et al., 1968). In this control methodology, the converter is described as a piece wise linear system (as in Eq. (1)) and the control objective is to maintain voltage confined within a pre-specified region, given two ad-doc surfaces. For example, in buck converters such surfaces will be: 0 vVS refoff (12) 0 vVS refon For boost or fly-back converters (among others), current limitation has also be performed in order to avoid system instability. Some options to avoid such instability are the use of a current dependent switching surface or some kind of multi-loop control. On the other hand, since in this kind of controller the control input (switching surfaces and assign mode criterion) does not depend on system parameters, such control strategy results very robust (model free controller). However, a drawback of this controller is that it may lead to variable frequency operation. Since the first works about hysteresis control (Babaa et al., 1968), various improvements have been performed. In (Leung & Chung, 2009), a time varying surface is proposed in order to reject large signal perturbation in two switching cycles. In (Schild et al., 2009) a procedure that uses both clock-driven and event-driven switching Hybrid Control of DC-DC Power Converters 187Renewable Energy186 that lead us to conclude asymptotic stability of a general HS may be hard to find. However, for switched systems such a task is easier in some cases.‡ In the following subsection we will describe some contributions in the area of hybrid control applied to DC-DC power converter. The discussion about their stability properties will be explained in view of the concepts and ideas stated here. 4. Types of HC in Power Electronics The types of HC found in the literature can be classified in two groups mainly; a first one uses either a switched description or a switching control action in the converter; in these papers, the control objective is to find a switching law such that voltage is tracked. A second group, which description of the converter may be continuous or piece wise continuous, in these strategies the control objective is aimed to solve optimally (time or using and index performance) the voltage regulation problem. We will refer to the first group as Hybrid Switched Control (HSC) and to the second as, Hybrid Optimal Control (HOBC). 4.1 Hybrid Switched Control In this category we can find variable structure controllers such as: hysteresis, sliding and current peak mode controllers among the most common. Sliding mode controllers (SMC) are usually designed based on an averaged continuous model of the converter. In general, the converter is described by non-linear differential equations evolving in a fictitious averaged state space. Under this situation the control input is also continuous but restricted to  1,0u . Once a continuous model is obtained, the next step is to define a suitable surface that resembles the desired steady state, usually such surface is a plane of the form: niii xaS1~ (11) where x~ =x-xref. and xref the desired state. However, the sliding surface can be also non-linear (Sira-Ramirez & Ríos-Bolivar, 1994), (Tokat et al., 2002). Sliding mode control action can be divided in two operation modes: the reaching mode and the sliding mode. The reaching mode is related with the operation phase in which from an arbitrary initial condition the trajectories goes toward the surface S. The control input of SMC can be seen also as divided in two, ueq that acts mostly on the reaching mode and ud for the sliding. The input ueq is usually continuous and parameter dependent, while ud is a discontinuous input usually depending on the sign of the desired surface therefore, is parameter insensitive. There exist various papers in the literature dealing with the robust design of controllers for the reaching mode. For example in (Sira-Ramirez & Ríos-Bolivar, 1994) a passivity based controller is proposed to enhance the robustness properties of SMC. The controller make use of the passive relation between input voltage and inductor current of the converter, along with ‡There exist trivial cases regarding stability of switched systems that are not of our interest. For example, if there is a stable mode in the subsystems family, we can guarantee stability of the switched system just by staying indefinitely on it, among others cases. Lyapunov based arguments to show that the desired tracking dynamics approaches the origin asymptotically (i.e. the equilibrium point is asymptotically stable). The use of non-linear and time-varying sliding surfaces has also been proposed to enhance the robustness of SMC (Sira-Ramirez & Ríos-Bolivar, 1994), (Sira-Ramirez et al., 1995), (Tokat et al., 2002). On the other hand, sensorless SMC has also been proposed, as well as some techniques for fixed frequency operation, the interested reader is refer to (Tan et al., 2008) for a comprehensive review of sliding mode control in power electronics. Fig. 4. Types of hybrid controllers used in DC-DC power converters. Hysteresis control can be interpreted as a switched control with two output-dependent switching surfaces: turn-on and turn-off surfaces (Babaa et al., 1968). In this control methodology, the converter is described as a piece wise linear system (as in Eq. (1)) and the control objective is to maintain voltage confined within a pre-specified region, given two ad-doc surfaces. For example, in buck converters such surfaces will be: 0 vVS refoff (12) 0 vVS refon For boost or fly-back converters (among others), current limitation has also be performed in order to avoid system instability. Some options to avoid such instability are the use of a current dependent switching surface or some kind of multi-loop control. On the other hand, since in this kind of controller the control input (switching surfaces and assign mode criterion) does not depend on system parameters, such control strategy results very robust (model free controller). However, a drawback of this controller is that it may lead to variable frequency operation. Since the first works about hysteresis control (Babaa et al., 1968), various improvements have been performed. In (Leung & Chung, 2009), a time varying surface is proposed in order to reject large signal perturbation in two switching cycles. In (Schild et al., 2009) a procedure that uses both clock-driven and event-driven switching Hybrid Control of DC-DC Power Converters 187Renewable Energy188 criterion for controlling the stationary operation of a converter is proposed. The switching surfaces are planes that can be systematically computed using periodic control systems. Consequently, desired loop properties such as orbital stability of a limit cycle and a fast transient response are guaranteed at least in a local neighbourhood around a nominal set point. In (Tao et al., 2006) variable hysteresis band control is proposed to achieve zero voltage switching for a multiple input converter. On the other hand, in (Kimball, 2006) a sensorless hysteresis-type controller is proposed. Such strategy does not need current measurements and it can be implemented in converters operating under DCM conditions. In (Khaligh & Emadi, 2007) a hysteresis type controller that works under fixed frequency switching conditions is proposed. At this point, it is interesting to notice that although most hysteresis controllers have parameter insensitivity§ and their stability and performance have been evaluated experimentally (see for example (Hwu & Yau, 2004); there exist a little formal theoretical support that explain their robust capabilities. Formal results involving robust practical stability of buck and boost converters under hysteresis-like controllers have been presented in (Cervantes et al., 2009a), (Cervantes et al., 2009b), (Perez-Pinal & Cervantes, 2009b); however, there are a lot of questions that remain open, specially for time varying and non-linear hysteresis (switching) surfaces. Peak Current Mode Control (PCMC) is a class of switched control that has both time dependent and state dependent surfaces. This control strategy was conceived to limit inductor current while guaranteeing a fixed frequency switching operation. For a buck converter the PCMC is the following If i> imax then switch goes off (13) If TTt  ),mod(0 then switch goes on (14) For PMCM to be well defined it is necessary that the switching condition (12) be of higher priority than the one in (11). It is well known that, for duty cycles greater than 50%, PCMC leads to converter instability and strange behaviour such as bifurcation, chaos (Barnejee & Verghese, 2001) and breathing** (Cervantes & Femat, 2008). However, a stabilizing ramp is used to avoid such undesirable behaviour. Such a ramp makes a switching surface be both state and time dependent (the one that limits current). In spite of the well-known capabilities of PMCM, as in the case of hysteresis control, formal studies regarding stability are rare; hence issues regarding local or global stability capabilities are not yet completely known, as well as robustness properties of the controller. There exist other expressions of switching control in the literature, for example the so-called “boundary” †† control studied in (Leung & Cheun, 2004), (Song & Chung, 2008) and (Leung & § Due to this fact, hysteresis control can be applied directly in many topologies without need of adjustment. This control is also called by some authors universal (Khaligh & Emadi, 2007). ** If a low frequency variable voltage input is used. ††This term is rarely used since confusion arise with the so-called boundary control in distributed system (infinite order systems) in which the frontier or boundary condition is Cheun, 2007). Such controllers use a piece-wise linear description of the converter and switching surfaces which can be either planes or second order surfaces. It is shown that second order surfaces display better transient behaviour than planes, however the implementation of such surfaces may be quite involved. The design of the surfaces is based on one hand, on the regulation requirements of the converter; and on the other, on natural trajectories of the system, that reduce the transient response of the systems (ideally in two switching actions). The switching surfaces depend on system parameters and although a sensitivity study is performed in (Leung & Cheun, 2007) with respect to system inductance and capacitance, the surfaces depend dramatically on load resistance; such fact may limit the advantages of the controller if such parameters are uncertain. Computation of switching instants based on an energy balance of the converter is shown in (Gupta & Patra, 2005). In this controller, the modes in which the converter gains and looses energy are identified. The switching instants are chosen so the total energy change is cero. Although this strategy has a physical motivation, the proposed strategy has the disadvantage of leading to variable current ripple and variable frequency operation. Another expression of switching control can be found in (Senesky et al., 2003). In such a reference, switched systems are described as hybrid automata. Based on this description the stabilization problem is stated in terms of the existence (and design) of suitable guards that guarantee that converter states remain in a safe set. In this excellent work, two regulation strategies are designed, the minimum ripple control and the minimum switching control that guarantee that a set of interest is controlled invariant (i.e. the system is practically stable). The proposed strategies have the advantage of having rigorous mathematical support and the effect of sample and perturbation is analysed. Although such techniques are probably easy to implement, unfortunately the authors do not show any experimental evidence of controller performance. On the other hand, other controllers based on the hybrid automata can be found (Sreekumar & Agarwal, 2007) (Sreekumar & Agarwal, 2008). In such works, a controller that is able to operate under both CCM and DCM is introduced. The authors give experimental evidence of the controller performance but unfortunately, the stability analysis presented do not take into account dynamic behaviour. 4.2 Hybrid Optimal Control As pointed out before, in HOC the description of the converter is continuous or piece wise continuous and the control strategy is aimed to solve an optimal problem. Within this category we find predictive control of converters e.g. (Geyer et al., 2008), (Morari et al., 2006). In this case, the control is usually operated in fixed frequency and the optimization problem can be solved: i) once every switching cycle, ii) every fixed number of switching cycles or iii) offline. The optimization problem is usually stated, based on a performance index function that penalizes voltage error tracking, control deviation from nominal, among other common criteria. Major concerns when dealing with optimization problems are the solution existence and its computation time. Due to these facts, usually some simplifications to the optimal problem are performed and off-line computation of the control law is preferred. manipulated in order to control variables that are spatial dependent (You & Lee, 1989) (Morgul, 1992). Hybrid Control of DC-DC Power Converters 189Renewable Energy188 criterion for controlling the stationary operation of a converter is proposed. The switching surfaces are planes that can be systematically computed using periodic control systems. Consequently, desired loop properties such as orbital stability of a limit cycle and a fast transient response are guaranteed at least in a local neighbourhood around a nominal set point. In (Tao et al., 2006) variable hysteresis band control is proposed to achieve zero voltage switching for a multiple input converter. On the other hand, in (Kimball, 2006) a sensorless hysteresis-type controller is proposed. Such strategy does not need current measurements and it can be implemented in converters operating under DCM conditions. In (Khaligh & Emadi, 2007) a hysteresis type controller that works under fixed frequency switching conditions is proposed. At this point, it is interesting to notice that although most hysteresis controllers have parameter insensitivity§ and their stability and performance have been evaluated experimentally (see for example (Hwu & Yau, 2004); there exist a little formal theoretical support that explain their robust capabilities. Formal results involving robust practical stability of buck and boost converters under hysteresis-like controllers have been presented in (Cervantes et al., 2009a), (Cervantes et al., 2009b), (Perez-Pinal & Cervantes, 2009b); however, there are a lot of questions that remain open, specially for time varying and non-linear hysteresis (switching) surfaces. Peak Current Mode Control (PCMC) is a class of switched control that has both time dependent and state dependent surfaces. This control strategy was conceived to limit inductor current while guaranteeing a fixed frequency switching operation. For a buck converter the PCMC is the following If i> imax then switch goes off (13) If TTt  ),mod(0 then switch goes on (14) For PMCM to be well defined it is necessary that the switching condition (12) be of higher priority than the one in (11). It is well known that, for duty cycles greater than 50%, PCMC leads to converter instability and strange behaviour such as bifurcation, chaos (Barnejee & Verghese, 2001) and breathing** (Cervantes & Femat, 2008). However, a stabilizing ramp is used to avoid such undesirable behaviour. Such a ramp makes a switching surface be both state and time dependent (the one that limits current). In spite of the well-known capabilities of PMCM, as in the case of hysteresis control, formal studies regarding stability are rare; hence issues regarding local or global stability capabilities are not yet completely known, as well as robustness properties of the controller. There exist other expressions of switching control in the literature, for example the so-called “boundary” †† control studied in (Leung & Cheun, 2004), (Song & Chung, 2008) and (Leung & § Due to this fact, hysteresis control can be applied directly in many topologies without need of adjustment. This control is also called by some authors universal (Khaligh & Emadi, 2007). ** If a low frequency variable voltage input is used. ††This term is rarely used since confusion arise with the so-called boundary control in distributed system (infinite order systems) in which the frontier or boundary condition is Cheun, 2007). Such controllers use a piece-wise linear description of the converter and switching surfaces which can be either planes or second order surfaces. It is shown that second order surfaces display better transient behaviour than planes, however the implementation of such surfaces may be quite involved. The design of the surfaces is based on one hand, on the regulation requirements of the converter; and on the other, on natural trajectories of the system, that reduce the transient response of the systems (ideally in two switching actions). The switching surfaces depend on system parameters and although a sensitivity study is performed in (Leung & Cheun, 2007) with respect to system inductance and capacitance, the surfaces depend dramatically on load resistance; such fact may limit the advantages of the controller if such parameters are uncertain. Computation of switching instants based on an energy balance of the converter is shown in (Gupta & Patra, 2005). In this controller, the modes in which the converter gains and looses energy are identified. The switching instants are chosen so the total energy change is cero. Although this strategy has a physical motivation, the proposed strategy has the disadvantage of leading to variable current ripple and variable frequency operation. Another expression of switching control can be found in (Senesky et al., 2003). In such a reference, switched systems are described as hybrid automata. Based on this description the stabilization problem is stated in terms of the existence (and design) of suitable guards that guarantee that converter states remain in a safe set. In this excellent work, two regulation strategies are designed, the minimum ripple control and the minimum switching control that guarantee that a set of interest is controlled invariant (i.e. the system is practically stable). The proposed strategies have the advantage of having rigorous mathematical support and the effect of sample and perturbation is analysed. Although such techniques are probably easy to implement, unfortunately the authors do not show any experimental evidence of controller performance. On the other hand, other controllers based on the hybrid automata can be found (Sreekumar & Agarwal, 2007) (Sreekumar & Agarwal, 2008). In such works, a controller that is able to operate under both CCM and DCM is introduced. The authors give experimental evidence of the controller performance but unfortunately, the stability analysis presented do not take into account dynamic behaviour. 4.2 Hybrid Optimal Control As pointed out before, in HOC the description of the converter is continuous or piece wise continuous and the control strategy is aimed to solve an optimal problem. Within this category we find predictive control of converters e.g. (Geyer et al., 2008), (Morari et al., 2006). In this case, the control is usually operated in fixed frequency and the optimization problem can be solved: i) once every switching cycle, ii) every fixed number of switching cycles or iii) offline. The optimization problem is usually stated, based on a performance index function that penalizes voltage error tracking, control deviation from nominal, among other common criteria. Major concerns when dealing with optimization problems are the solution existence and its computation time. Due to these facts, usually some simplifications to the optimal problem are performed and off-line computation of the control law is preferred. manipulated in order to control variables that are spatial dependent (You & Lee, 1989) (Morgul, 1992). Hybrid Control of DC-DC Power Converters 189Renewable Energy190 In (Geyer et al., 2008) a predictive control of a step-down converter is proposed. The controller makes use of a piece-wise continuous model that is used to generate a discrete time approach of the system in order to simplify the solution of the constrained optimal control. The optimization problem is solved offline using dynamic programming and a Kalman filter is used to ensure voltage tracking and load estimation. The stability analysis is performed using piece-wise Lyapunov functions to show that a discrete equilibrium point is exponentially stable. Observe that in contrast to the discussed in Section 3, exponential stability in the discrete case can be guarantee since a discrete equilibrium point can be either i) a constant value or ii) a periodic orbit (which is the case in converters). A similar predictive control is also proposed in (Almer et al., 2007) but in this case, a multi-loop robust controller that uses discrete integral actions to compensate uncertainty and track the reference voltage is used. On the other hand, an optimum time controller has been proposed in (Chen et al., 2006) for a series-parallel resonant converter. In this paper a piece-wise continuous model of the converter is used to analyze the state-state trajectories to compute optimal switching instants. The optimal control is able to compensate step voltage commands theoretically in one switching cycle. The controller makes use of voltage and current measurements as well as zero voltage crossing detection devices. Although the authors show the experimental performance of the controller, stability analysis is not complete and robustness of the proposed control has to be studied. It can be observed that HOC can be much difficult to synthesize than switched controllers, such fact may affect adversely the complexity of actual implementations. Since these controllers are much complex, their application has to be well justified in terms of performance improvement or robust stability capabilities. 4.3 High Order Converter Hybrid control and in particular, switched control has been gained more attention lately, among other things, due to its property of being easily implemented even in high order converters (interleaved or cascaded converters). These types of converters are mainly used in medium to high power applications in order to divide the converter power in cells, allowing managing lower currents and faster switching devices. In some topologies, there is a dramatic increasing in the number of switches. Incrementing the number of switches, increment also the number of degrees of freedom of the system, which in some cases may affect adversely the complexity of controller design. However, also open the possibility to solve more efficiently some operation problems. For example, in interleaved converters, the possibility of reducing substantially voltage rippling arises even at low switching frequencies. Moreover, perturbation rejection can be performed smoothly, since input actions and efforts can be divided among the branches (Batchvarov et al., 2000). In spite of these advantages, there exists in the literature very few works leading with voltage regulation and ripple limitation in higher order DC-DC power converters. In (Baja et al., 2008) a stabilizing approach used to control capacitors voltage and current of a four-level three-cell DC-DC converter is proposed. The description of the converter is performed using bond graph formalism‡‡. Under the assumption that switching actions are performed in pairs, the notion of commutation cell is used to allow the converter description via bond graph formalism be topologic simple and with fixed causality in any mode. Three types of cells are detected: source, internal and load cells, for each one a static model is derived that relates inputs and outputs. The dynamic description of the converter takes also into account the dynamics of the Boolean control variables. The controller design is based on Lyapunov redesign technique and local atractivity of system trajectories is guarantee to a neighborhood of the origin. On the other hand, a hysteresis like controller is proposed in (Cervantes et al., 2009b) for the case of interleaved converters. The controlled derivation is based on an approximated switched description of the system. Closed-loop stability analysis of the converter is performed leading to a practical stability result. The proposed controller is able to regulate voltage, while warranting maximum current and voltage rippling, as well as current balancing. The performance of the controller is tested via both numerical simulations and experimental work; and it is shown that the controller can be readily implemented using integrated circuits. In multi cell converters, the number of dynamic elements increases compared to traditional converters; therefore, it is convenient to use sensor-less control approaches, otherwise, the cost of the control implementation may increase dramatically. After a short review of the literature and in spite of the importance of developing such controllers, we could establish a truly need of investigation of this subject in the case of hybrid controllers. Furthermore, stability analysis of switching control in both parallel and cascaded topologies is also an open problem, as well as the development of a systematic procedure to design switching surfaces that lead to the solution of highly versatile control specifications. 5. Conclusions In this work, a general description and classification of the most common hybrid controllers for DC-DC power converters is given. It is stated that in particular switched control has the advantage that some performance requirements can be formulated explicitly by defining suitable switching surfaces. Moreover, in some cases, the switching criterion and the mode assignment criterion can be manipulated independently to obtain multiple control input schemes. A main advantage of switched control is its inherent flexibility to operate in a wide operation region. Hybrid control also represents an option to solve optimal operation problems, for example, minimum switching control or predictive-type control. Furthermore, hybrid controllers have been designed to solve voltage regulation problems even if the converter changes from CCM to DCM regime and vice versa. Although some ‡‡ Bond graph modeling is a method for describing dynamic systems, which is based on the assumption that it is possible to define the characteristics of the subsystems and the connections between them without energy losses. Bond graph formalism may be perceived some how close to graph automata description, however there are substantial differences, see (Antic et al., 1999) for more details. Hybrid Control of DC-DC Power Converters 191Renewable Energy190 In (Geyer et al., 2008) a predictive control of a step-down converter is proposed. The controller makes use of a piece-wise continuous model that is used to generate a discrete time approach of the system in order to simplify the solution of the constrained optimal control. The optimization problem is solved offline using dynamic programming and a Kalman filter is used to ensure voltage tracking and load estimation. The stability analysis is performed using piece-wise Lyapunov functions to show that a discrete equilibrium point is exponentially stable. Observe that in contrast to the discussed in Section 3, exponential stability in the discrete case can be guarantee since a discrete equilibrium point can be either i) a constant value or ii) a periodic orbit (which is the case in converters). A similar predictive control is also proposed in (Almer et al., 2007) but in this case, a multi-loop robust controller that uses discrete integral actions to compensate uncertainty and track the reference voltage is used. On the other hand, an optimum time controller has been proposed in (Chen et al., 2006) for a series-parallel resonant converter. In this paper a piece-wise continuous model of the converter is used to analyze the state-state trajectories to compute optimal switching instants. The optimal control is able to compensate step voltage commands theoretically in one switching cycle. The controller makes use of voltage and current measurements as well as zero voltage crossing detection devices. Although the authors show the experimental performance of the controller, stability analysis is not complete and robustness of the proposed control has to be studied. It can be observed that HOC can be much difficult to synthesize than switched controllers, such fact may affect adversely the complexity of actual implementations. Since these controllers are much complex, their application has to be well justified in terms of performance improvement or robust stability capabilities. 4.3 High Order Converter Hybrid control and in particular, switched control has been gained more attention lately, among other things, due to its property of being easily implemented even in high order converters (interleaved or cascaded converters). These types of converters are mainly used in medium to high power applications in order to divide the converter power in cells, allowing managing lower currents and faster switching devices. In some topologies, there is a dramatic increasing in the number of switches. Incrementing the number of switches, increment also the number of degrees of freedom of the system, which in some cases may affect adversely the complexity of controller design. However, also open the possibility to solve more efficiently some operation problems. For example, in interleaved converters, the possibility of reducing substantially voltage rippling arises even at low switching frequencies. Moreover, perturbation rejection can be performed smoothly, since input actions and efforts can be divided among the branches (Batchvarov et al., 2000). In spite of these advantages, there exists in the literature very few works leading with voltage regulation and ripple limitation in higher order DC-DC power converters. In (Baja et al., 2008) a stabilizing approach used to control capacitors voltage and current of a four-level three-cell DC-DC converter is proposed. The description of the converter is performed using bond graph formalism‡‡. Under the assumption that switching actions are performed in pairs, the notion of commutation cell is used to allow the converter description via bond graph formalism be topologic simple and with fixed causality in any mode. Three types of cells are detected: source, internal and load cells, for each one a static model is derived that relates inputs and outputs. The dynamic description of the converter takes also into account the dynamics of the Boolean control variables. The controller design is based on Lyapunov redesign technique and local atractivity of system trajectories is guarantee to a neighborhood of the origin. On the other hand, a hysteresis like controller is proposed in (Cervantes et al., 2009b) for the case of interleaved converters. The controlled derivation is based on an approximated switched description of the system. Closed-loop stability analysis of the converter is performed leading to a practical stability result. The proposed controller is able to regulate voltage, while warranting maximum current and voltage rippling, as well as current balancing. The performance of the controller is tested via both numerical simulations and experimental work; and it is shown that the controller can be readily implemented using integrated circuits. In multi cell converters, the number of dynamic elements increases compared to traditional converters; therefore, it is convenient to use sensor-less control approaches, otherwise, the cost of the control implementation may increase dramatically. After a short review of the literature and in spite of the importance of developing such controllers, we could establish a truly need of investigation of this subject in the case of hybrid controllers. Furthermore, stability analysis of switching control in both parallel and cascaded topologies is also an open problem, as well as the development of a systematic procedure to design switching surfaces that lead to the solution of highly versatile control specifications. 5. Conclusions In this work, a general description and classification of the most common hybrid controllers for DC-DC power converters is given. It is stated that in particular switched control has the advantage that some performance requirements can be formulated explicitly by defining suitable switching surfaces. Moreover, in some cases, the switching criterion and the mode assignment criterion can be manipulated independently to obtain multiple control input schemes. A main advantage of switched control is its inherent flexibility to operate in a wide operation region. Hybrid control also represents an option to solve optimal operation problems, for example, minimum switching control or predictive-type control. Furthermore, hybrid controllers have been designed to solve voltage regulation problems even if the converter changes from CCM to DCM regime and vice versa. Although some ‡‡ Bond graph modeling is a method for describing dynamic systems, which is based on the assumption that it is possible to define the characteristics of the subsystems and the connections between them without energy losses. Bond graph formalism may be perceived some how close to graph automata description, however there are substantial differences, see (Antic et al., 1999) for more details. Hybrid Control of DC-DC Power Converters 191Renewable Energy192 HC methodologies are designed to operate at varying frequency (mostly transient response), there is an increasing tendency in the literature to investigate fixed frequency hybrid control schemes. There are still a lot of open questions regarding stability and design of HC. We find a research opportunity in the formal stability study of some of the proposed control methodologies, in order to reveal advantages and limitations. We also find a research opportunity in the systematic design of switching laws based on time-varying nonlinear switching surfaces as well as on the design of sensor-less hybrid control of higher order converters. Summarizing, hybrid controllers represent a versatile solution to guarantee energy quality in the renewable energy field, but the uncontroversial success of this control methodology will reside in the development of systematic approaches that guarantee both, simplicity in the implementation and robust stability properties. 6. References Aeyels D. & Peuteman J. (1998). A new asymptotic stability criterion for nonlinear time-variant differential equations, IEEE Trans. Automat. Contr., vol. 43, no. 7, pp. 968–971, July 1998, ISSN 0018-9286. Almer S., Fujioka H. Jonsson U, Kao C.Y., Patino D., Riedinger R., Geyer T., Buccuti A., Papafotiou G., Morari M., Wernrud A & Rantzer A., (2007). Hybrid control techniques for switched-mode DC-DC converters. Part I. The step-down topology. Proc. of American Control Conference, pp. 5450-5457, ISBN 1-4244-0989-6, July11-13, New York. Antic, D.; Vidojkovic, B.; Maldenovic, M. (1999). An introduction to bond graph modelling of dynamic systems, Proc. of Telecommunications in Modern Satellite, Cable and Broadcasting Services Conference, 1999. Vol. 2, No. 13-15 Oct. 1999 Page(s):661 – 664. Austria-Gonzalez, L.F., Perez-Pinal, F.J. & Cervantes I.(2008). Hybrid Multi-objective Control of DC-DC. Proc. of IEEE Vehicle Power and Propulsion Conference. IEEE Harbin, China, República Popular. septiembre 2008. Babaa, I.; Wilson, T. & Yuan Y. (1968). Analytic solutions of limit cycles in a feedback-regulated converter system with hysteresis; IEEE Trans. Automatic Control, Vol. 13, No. 5 Oct 1968 Page(s): 524 – 531, ISSN 0018-9286. Baja, M., Cormerais H. & Buisson J. (2008). Modeling and Hybrid control of a four-level three cell DC-DC converter. Proc of Industrial Electronics Conf. 10-13 Nov. 2008, Page(s): 3278-3283. ISBN: 978-1-4244-1766-7 Barnejee, S. & Vergese G.C. (2001). Nonlinear Phenomena in Power Electronics. Attractors, Bifurcations, Chaos, and Nonlinear Control, Ed. IEEE Press, ISBN 0-7803-5883-8, New York. Batchvarov, J.S.; Duarte, J.L. & Hendrix, M.A.M. (2000). Interleaved converters based on hysteresis current control; Proc. of Power Electronics Specialists Conference, 2000. PESC 00. 2000 Vol. 2, 18-23 June 2000 pp: 655 – 661. Branicky, M.S. (1998). Multiple Lyapunov functions and other analysis tools for switched and hybrid systems, IEEE Trans. Automat. Contr., vol. 43, pp. 1679–1684, 1998, ISSN 0018-9286. Capel A., Ferrante G., et al., (1978) ”Application of injected current mode control for the dynamic analysis of switching regulators with new concept of LC3 modulator,” IEEE PESC Record, pp.135-147. Cervantes I., Femat R. (2008). Intermittence in Linear Driven Switched Systems. Int. J. Bif. Chaos, vol. 18, pp. 495-508. Cervantes, I. Mendoza-Torres, A. Garcia-Cuevas, R. & Perez-Pinal. J. (2009b) Switched Control of Interleaved Converters. Proc. of IEEE Int. Vehicle Power and Propulsion Conference. Dearborn, Michigan, September 7-11, 2009. Cervantes, I., Mendoza-Torres, A. & Perez-Pinal. J. (2009a) High Performance Hybrid Control for Buck Converters. Proc. 13th European Conference on Power Electronics and Applications - EPE 2009 8-10 September 2009, Barcelona, Spain. Chen H., Sng, E.K.K & Tseng K.J.T (2006). Optimum trajectory switching control for series-parallel resonant converter. IEEE. Trans. Ind. Electronics, vol. 53, No. 5, pp. 1555-1563, ISSN 0278-0046. Deane J. H. B. and Hamill D. C. (1990) Instability, Subharmonics, and Chaos in Power Electronic Systems, IEEE Transactions on Power Electronics, 260-268, ISSN: 0885-8993. ECPE, European Center for Power Electronics, EPE European Power Electronics and Drives Association (2007), “Position Paper on Energy Efficiency – the Role of Power Electronics,” March. http://www.ecpe.org/download/power_electronic/Position _Paper_Energy_Efficiency.pdf, accessed February 1st 2009. Enerdata, (2008), “The world energy demand in 2007,” pp. 1-17 Enerdata, (2008b), “2020 Global Energy Forecasts”, pp. 1-21. Erickson R. & Maksimovic D. (2001), ”Principles of Power Electronics,” Kluwer Academic Publishers. Erickson R., Cuk S., and Middlebrook R. (1982), ”Large signal modeling and analysis of switching regulators,” Proc. of IEEE Power Electronics Specialists Conference Records, pp. 240-250 Filipov, A.F. (1998). Differential Equations with Discontinuous Right-Hands Sides, Ed. Arscott, F.M., Kluwer Academic Publishers ISBN. 902772699X. Geyer T., Papafotiou G. & Morari M. (2008). Hybrid model predictive control of the step-down DC-DC converter. IEEE Trans. Control System Technology, vol 16, No. 6, pp. 1112- 1124, ISSN. 1063-6536. Goebel, R., Sanfelice R.G. & Teel, A.R. (2009). Hybrid dynamical systems. IEEE Control Systems Magazine, pp. 28-93, ISSN. 066-033x Gupta, P. & Patra A.N. (2005) Super stable energy based switching control scheme for DC-DC buck converter circuits. Proc of IEEE Int.Symp on Circuts and Systems. pp. 3063-3066, ISBN 0-7803-8834-8 Henzinger, T. A. (1996). The Theory of Hybrid Automata, Proc.11th Annual Symposium on Logic in Computer Science, 278-292. IEEE Computer Society Press Henzinger, T.A. (1995). Hybrid Automata with Finite Bisimulations. ICALP 95: Automata, Languages and Programming, Lecture Notes in Computer Science, 944, pp 225-238. Springer Verlag. Hybrid Control of DC-DC Power Converters 193Renewable Energy192 HC methodologies are designed to operate at varying frequency (mostly transient response), there is an increasing tendency in the literature to investigate fixed frequency hybrid control schemes. There are still a lot of open questions regarding stability and design of HC. We find a research opportunity in the formal stability study of some of the proposed control methodologies, in order to reveal advantages and limitations. We also find a research opportunity in the systematic design of switching laws based on time-varying nonlinear switching surfaces as well as on the design of sensor-less hybrid control of higher order converters. Summarizing, hybrid controllers represent a versatile solution to guarantee energy quality in the renewable energy field, but the uncontroversial success of this control methodology will reside in the development of systematic approaches that guarantee both, simplicity in the implementation and robust stability properties. 6. References Aeyels D. & Peuteman J. (1998). A new asymptotic stability criterion for nonlinear time-variant differential equations, IEEE Trans. Automat. Contr., vol. 43, no. 7, pp. 968–971, July 1998, ISSN 0018-9286. Almer S., Fujioka H. Jonsson U, Kao C.Y., Patino D., Riedinger R., Geyer T., Buccuti A., Papafotiou G., Morari M., Wernrud A & Rantzer A., (2007). Hybrid control techniques for switched-mode DC-DC converters. Part I. The step-down topology. Proc. of American Control Conference, pp. 5450-5457, ISBN 1-4244-0989-6, July11-13, New York. Antic, D.; Vidojkovic, B.; Maldenovic, M. (1999). An introduction to bond graph modelling of dynamic systems, Proc. of Telecommunications in Modern Satellite, Cable and Broadcasting Services Conference, 1999. Vol. 2, No. 13-15 Oct. 1999 Page(s):661 – 664. Austria-Gonzalez, L.F., Perez-Pinal, F.J. & Cervantes I.(2008). Hybrid Multi-objective Control of DC-DC. Proc. of IEEE Vehicle Power and Propulsion Conference. IEEE Harbin, China, República Popular. septiembre 2008. Babaa, I.; Wilson, T. & Yuan Y. (1968). Analytic solutions of limit cycles in a feedback-regulated converter system with hysteresis; IEEE Trans. Automatic Control, Vol. 13, No. 5 Oct 1968 Page(s): 524 – 531, ISSN 0018-9286. Baja, M., Cormerais H. & Buisson J. (2008). Modeling and Hybrid control of a four-level three cell DC-DC converter. Proc of Industrial Electronics Conf. 10-13 Nov. 2008, Page(s): 3278-3283. ISBN: 978-1-4244-1766-7 Barnejee, S. & Vergese G.C. (2001). Nonlinear Phenomena in Power Electronics. Attractors, Bifurcations, Chaos, and Nonlinear Control, Ed. IEEE Press, ISBN 0-7803-5883-8, New York. Batchvarov, J.S.; Duarte, J.L. & Hendrix, M.A.M. (2000). Interleaved converters based on hysteresis current control; Proc. of Power Electronics Specialists Conference, 2000. PESC 00. 2000 Vol. 2, 18-23 June 2000 pp: 655 – 661. Branicky, M.S. (1998). Multiple Lyapunov functions and other analysis tools for switched and hybrid systems, IEEE Trans. Automat. Contr., vol. 43, pp. 1679–1684, 1998, ISSN 0018-9286. Capel A., Ferrante G., et al., (1978) ”Application of injected current mode control for the dynamic analysis of switching regulators with new concept of LC3 modulator,” IEEE PESC Record, pp.135-147. Cervantes I., Femat R. (2008). Intermittence in Linear Driven Switched Systems. Int. J. Bif. Chaos, vol. 18, pp. 495-508. Cervantes, I. Mendoza-Torres, A. Garcia-Cuevas, R. & Perez-Pinal. J. (2009b) Switched Control of Interleaved Converters. Proc. of IEEE Int. Vehicle Power and Propulsion Conference. Dearborn, Michigan, September 7-11, 2009. Cervantes, I., Mendoza-Torres, A. & Perez-Pinal. J. (2009a) High Performance Hybrid Control for Buck Converters. Proc. 13th European Conference on Power Electronics and Applications - EPE 2009 8-10 September 2009, Barcelona, Spain. Chen H., Sng, E.K.K & Tseng K.J.T (2006). Optimum trajectory switching control for series-parallel resonant converter. IEEE. Trans. Ind. Electronics, vol. 53, No. 5, pp. 1555-1563, ISSN 0278-0046. Deane J. H. B. and Hamill D. C. (1990) Instability, Subharmonics, and Chaos in Power Electronic Systems, IEEE Transactions on Power Electronics, 260-268, ISSN: 0885-8993. ECPE, European Center for Power Electronics, EPE European Power Electronics and Drives Association (2007), “Position Paper on Energy Efficiency – the Role of Power Electronics,” March. http://www.ecpe.org/download/power_electronic/Position _Paper_Energy_Efficiency.pdf, accessed February 1st 2009. Enerdata, (2008), “The world energy demand in 2007,” pp. 1-17 Enerdata, (2008b), “2020 Global Energy Forecasts”, pp. 1-21. Erickson R. & Maksimovic D. (2001), ”Principles of Power Electronics,” Kluwer Academic Publishers. Erickson R., Cuk S., and Middlebrook R. (1982), ”Large signal modeling and analysis of switching regulators,” Proc. of IEEE Power Electronics Specialists Conference Records, pp. 240-250 Filipov, A.F. (1998). Differential Equations with Discontinuous Right-Hands Sides, Ed. Arscott, F.M., Kluwer Academic Publishers ISBN. 902772699X. Geyer T., Papafotiou G. & Morari M. (2008). Hybrid model predictive control of the step-down DC-DC converter. IEEE Trans. Control System Technology, vol 16, No. 6, pp. 1112- 1124, ISSN. 1063-6536. Goebel, R., Sanfelice R.G. & Teel, A.R. (2009). Hybrid dynamical systems. IEEE Control Systems Magazine, pp. 28-93, ISSN. 066-033x Gupta, P. & Patra A.N. (2005) Super stable energy based switching control scheme for DC-DC buck converter circuits. Proc of IEEE Int.Symp on Circuts and Systems. pp. 3063-3066, ISBN 0-7803-8834-8 Henzinger, T. A. (1996). The Theory of Hybrid Automata, Proc.11th Annual Symposium on Logic in Computer Science, 278-292. IEEE Computer Society Press Henzinger, T.A. (1995). Hybrid Automata with Finite Bisimulations. ICALP 95: Automata, Languages and Programming, Lecture Notes in Computer Science, 944, pp 225-238. Springer Verlag. Hybrid Control of DC-DC Power Converters 193Renewable Energy194 Hwu, K.I. & Yau Y.T. (2004). Perfomance improvement of hysteresis voltage-controlld forward converters using FPGA-based SR control. Proc. of Int. Conf. Power System Technology, Singapure 21-24 Nov, 2004. pp. 281-287, ISBN 07803-8610-8. International Energy Outlook 2008, Energy Information Administration, US Government, Report#:DOE/EIA-0484(2008) Release Date: June. Khaligh A. & Emadi A. (2007). Suitability of the pulse adjustment technique to control single DC/DC choppers feeding vehicular constant power loads in parallel with conventional loads. Int. J. Electric and Hybrid Vehicles. Vol. 1, No., 1, 2007. Khalik H. S. (2002). Nonlinear Systems. Prentice Hall, ISBN 0-13-067389-7, New Jersey. Kimball, J.W.; Krein, P.T. & Chen Y. (2006). Hysteresis and delta modulation control of converters using sensorless current mode; IEEE Trans. Power Electronics, Vol.21, Issue 4, July 2006 Page(s):1154 – 1158, ISSN: 0885-8993. Lazar, M.; Heemels, W.P.M.H.; Weiland, S.; Bemporad (2006) Stabilizing Model Predictive Control of Hybrid Systems Lazar, IEEE Trans. Automatic Control, Vol. 51, No. 11, pp: 1813 – 1818, ISSN 0018-9286. Leung, K.K.-S. & Chung, H.S-H. (2007). A comparative study of boundary control with first and second order switching surfaces for buck converters opertin in DCM. IEEE Trans. Power Electr. Vol. 22, No. 4, pp. 1196-1209 ISSN: 0885-8993. Leung, K.K.S. & Chung, H.S.H. (2004) Derivation of a second-order switching surface in the boundary control of buck converters. IEEE Power Electronics Letters, Vol. 2, No. 2, Page numbers (63 – 67). Leung, K.K-S & Chung H. S.-H. (2009) Dynamic hysteresis band control of the buck converter with fast transient response; IEEE Trans. Circuits and Systems-II, Vol. 52, No 7, July 2005 Page(s):398 – 402 Li, Z; Soh, Y. & Wen, C. (2005). Switched and Impulsive Systems. Analysis, Design and Applications, Springer-Verlag, ISBN 3-540-23952-9, Berlin. Liberzon, D. (2003). Switching in Systems and Control, Birkhausen, ISBN 0-8176-4297-8, Boston. Lin H. (2008). Hybrid Output Feedback Stabilization for LTI Systems With Single Output IEEE Trans. Aut. Control, Vol. 53, No 7, pp: 1736 – 1740, ISSN 0018-9286. Middlebrook R. (1985), Topics in multi-loop regulators and current mode programming, Proc. IEEE Power Electronics Specialists Conf. (PESC), pp. 716-732. Middlebrook R. Cuk S. (1976), A general unified approach to modeling switching-converter power stages, Proc. IEEE Power Electronics Specialists Conf. (PESC), pp. 18-34. Mohan N. T. Undelead and P. Robbins William (1989), Power Electronics: Converters, Applications and Design, New York: John Wiley. Morari, M., Papafotiou, G., Buisson, J. & De Schutter B. (2006). Final Report on Modelling Tools, Benchmarks and Control Methods. Project report Hybrid Control: Taming Heterogeneity and Complexity of Networked Embedded Systems, 13/03/2006, v. 1.1. Morgul, O. (1992). Dynamic boundary control of a Euler-Bernoulli beam. IEEE Trans. on Automatic Control, Vol. 37, No.5, pp. 639-642, ISSN 0018-9286. Perez-Pinal J. & Cervantes. I. (2009b). Multi-Objective Control for Cascaded Boost Converter with SingleActive Switch Proc. IEEE International Electric Machines and Drives Conference: IEMDC 09 Miami Florida Perez-Pinal J. , Cervantes I., Leyva, J. & L.H. Diaz (2009a) Control of single active switch cascade boost converter for fuel cell applications Proc. 13th European Conference on Power Electronics and Applications - EPE 2009 8-10 September 2009, Barcelona, Spain. Quaicoe, J. E. & Iqbal, M. T (2008). Advanced Boundary Control of Inverters Using the Natural Switching Surface: Normalized Geometrical Derivation Ordonez, IEEE Trans. Power Electronics, Vol. 23, No. 6, pp. 2915-2930, ISSN: 0885-8993. Rajeev, A., Henzinger T.A. & Sontag, E.D. (1996). Hybrid Systems III. Verification and Control. Lecture Notes in Computer Science, 3, Springer Verlag. ISBN-13: 978-3540611554. New York. Rashid M. (1988), Power Electronics: Circuits, Devices and Applications, Englewood Cliffs, NJ: Prentice Hall. Redl, R. & Sokal, N. O. (1985) Current-Mode Control, Five Different Types, Used With the Three Basic Classes of Power Converters: Small-Signal AC and Large-Signal DC Characterization, Stability Requirements, and Implementation of Practical Circuits, IEEE Power Electronics Specialists Conference - 1985 Record, pp. 771-785. Rodriges L. & How J.P. (2003). Observer-based control of piecewise-affine systems. Int. J Control, vol 76, No. 5, pp. 459-477. Schild, A.; Lunze, J.; Krupar, J.; Schwarz, W. (2009) Design of Generalized Hysteresis Controllers for DC–DC Switching Power Converters.; IEEE Trans. Power Electronics, Vol. 24, No. 1, Jan. 2009, Page(s): 138 – 146, ISSN: 0885-8993. Senesky, M., Eirea, G. & Koo, T.J. (2003). Hybrid Modeling and Control of Power Electronics, Hybrid systems: computation and control, Lecture Notes in Computer Science Eds. Maler, O. & Pnueli A. Springer, 2003, pp. 450-465. ISBN:3-540-00913-2. Sira-Ramirez H. & Rios-Bolivar, M. (1994). Sliding Mode Control of dc-to-dc Power Converters via Extended Linearization, IEEE Trans Circ Syst-I, Vol. 41, No. 10 pp. 652-662. Sira-Ramirez, H.; Ortega, R.; Perez-Moreno, R.; Garcia-Esteban, M (1995). A sliding mode controller-observer for DC-to-DC power converters: a passivity approach. Proc. of IEEE Conf. Decision and Control, Vol. 4 Date: 13-15 Dec 1995, Pages: 3379 - 3384. Song T.-T. & Chung, H.S-H. (2008). Boundary control of boost converters using state-energy plane. IEEE Trans. Power Electr. Vol. 23, No. 2, pp. 551563 ISSN: 0885-8993. Sontag E.D. (1998). Mathematical Control Theory. Deterministic Finite Dimensional Systems. Springer Verlag, ISBN 0-387-98489-5 Sreekumar C. & Agarwal V. (2007). Hybrid control approach for the output voltage regulation in buck type DC-DC converter, IET Electr. Power Appl, vol 1, No. 6, pp 897-906. Sreekumar C. & Agarwal V. (2008). A hybrid control algorithm for voltage regulation in DC-DC Boost Converter, IEEE Trans. Industrial Electronics, vol 55, No. 6, pp 2530-2538. ISSN 0278-0046. Sreekumar C. &, Agarwal V. (2006). A Circuit Theoretical Approach to Hybrid Mode Switching Control of a Pseudo CCM Boost Converter, Proc IEEE International Conference on Industrial Technology, pp. 791-795, ISBN 1-4244-0726-5 , Numbai, India, 15-17 Dec. 2006 Sun, Z. & Ge, S.S. (2005). Switched Linear Systems. Control and design, Springer-Verlag, ISBN 1-852333-893-8, London. Hybrid Control of DC-DC Power Converters 195Renewable Energy194 Hwu, K.I. & Yau Y.T. (2004). Perfomance improvement of hysteresis voltage-controlld forward converters using FPGA-based SR control. Proc. of Int. Conf. Power System Technology, Singapure 21-24 Nov, 2004. pp. 281-287, ISBN 07803-8610-8. International Energy Outlook 2008, Energy Information Administration, US Government, Report#:DOE/EIA-0484(2008) Release Date: June. Khaligh A. & Emadi A. (2007). Suitability of the pulse adjustment technique to control single DC/DC choppers feeding vehicular constant power loads in parallel with conventional loads. Int. J. Electric and Hybrid Vehicles. Vol. 1, No., 1, 2007. Khalik H. S. (2002). Nonlinear Systems. Prentice Hall, ISBN 0-13-067389-7, New Jersey. Kimball, J.W.; Krein, P.T. & Chen Y. (2006). Hysteresis and delta modulation control of converters using sensorless current mode; IEEE Trans. Power Electronics, Vol.21, Issue 4, July 2006 Page(s):1154 – 1158, ISSN: 0885-8993. Lazar, M.; Heemels, W.P.M.H.; Weiland, S.; Bemporad (2006) Stabilizing Model Predictive Control of Hybrid Systems Lazar, IEEE Trans. Automatic Control, Vol. 51, No. 11, pp: 1813 – 1818, ISSN 0018-9286. Leung, K.K.-S. & Chung, H.S-H. (2007). A comparative study of boundary control with first and second order switching surfaces for buck converters opertin in DCM. IEEE Trans. Power Electr. Vol. 22, No. 4, pp. 1196-1209 ISSN: 0885-8993. Leung, K.K.S. & Chung, H.S.H. (2004) Derivation of a second-order switching surface in the boundary control of buck converters. IEEE Power Electronics Letters, Vol. 2, No. 2, Page numbers (63 – 67). Leung, K.K-S & Chung H. S.-H. (2009) Dynamic hysteresis band control of the buck converter with fast transient response; IEEE Trans. Circuits and Systems-II, Vol. 52, No 7, July 2005 Page(s):398 – 402 Li, Z; Soh, Y. & Wen, C. (2005). Switched and Impulsive Systems. Analysis, Design and Applications, Springer-Verlag, ISBN 3-540-23952-9, Berlin. Liberzon, D. (2003). Switching in Systems and Control, Birkhausen, ISBN 0-8176-4297-8, Boston. Lin H. (2008). Hybrid Output Feedback Stabilization for LTI Systems With Single Output IEEE Trans. Aut. Control, Vol. 53, No 7, pp: 1736 – 1740, ISSN 0018-9286. Middlebrook R. (1985), Topics in multi-loop regulators and current mode programming, Proc. IEEE Power Electronics Specialists Conf. (PESC), pp. 716-732. Middlebrook R. Cuk S. (1976), A general unified approach to modeling switching-converter power stages, Proc. IEEE Power Electronics Specialists Conf. (PESC), pp. 18-34. Mohan N. T. Undelead and P. Robbins William (1989), Power Electronics: Converters, Applications and Design, New York: John Wiley. Morari, M., Papafotiou, G., Buisson, J. & De Schutter B. (2006). Final Report on Modelling Tools, Benchmarks and Control Methods. Project report Hybrid Control: Taming Heterogeneity and Complexity of Networked Embedded Systems, 13/03/2006, v. 1.1. Morgul, O. (1992). Dynamic boundary control of a Euler-Bernoulli beam. IEEE Trans. on Automatic Control, Vol. 37, No.5, pp. 639-642, ISSN 0018-9286. Perez-Pinal J. & Cervantes. I. (2009b). Multi-Objective Control for Cascaded Boost Converter with SingleActive Switch Proc. IEEE International Electric Machines and Drives Conference: IEMDC 09 Miami Florida Perez-Pinal J. , Cervantes I., Leyva, J. & L.H. Diaz (2009a) Control of single active switch cascade boost converter for fuel cell applications Proc. 13th European Conference on Power Electronics and Applications - EPE 2009 8-10 September 2009, Barcelona, Spain. Quaicoe, J. E. & Iqbal, M. T (2008). Advanced Boundary Control of Inverters Using the Natural Switching Surface: Normalized Geometrical Derivation Ordonez, IEEE Trans. Power Electronics, Vol. 23, No. 6, pp. 2915-2930, ISSN: 0885-8993. Rajeev, A., Henzinger T.A. & Sontag, E.D. (1996). Hybrid Systems III. Verification and Control. Lecture Notes in Computer Science, 3, Springer Verlag. ISBN-13: 978-3540611554. New York. Rashid M. (1988), Power Electronics: Circuits, Devices and Applications, Englewood Cliffs, NJ: Prentice Hall. Redl, R. & Sokal, N. O. (1985) Current-Mode Control, Five Different Types, Used With the Three Basic Classes of Power Converters: Small-Signal AC and Large-Signal DC Characterization, Stability Requirements, and Implementation of Practical Circuits, IEEE Power Electronics Specialists Conference - 1985 Record, pp. 771-785. Rodriges L. & How J.P. (2003). Observer-based control of piecewise-affine systems. Int. J Control, vol 76, No. 5, pp. 459-477. Schild, A.; Lunze, J.; Krupar, J.; Schwarz, W. (2009) Design of Generalized Hysteresis Controllers for DC–DC Switching Power Converters.; IEEE Trans. Power Electronics, Vol. 24, No. 1, Jan. 2009, Page(s): 138 – 146, ISSN: 0885-8993. Senesky, M., Eirea, G. & Koo, T.J. (2003). Hybrid Modeling and Control of Power Electronics, Hybrid systems: computation and control, Lecture Notes in Computer Science Eds. Maler, O. & Pnueli A. Springer, 2003, pp. 450-465. ISBN:3-540-00913-2. Sira-Ramirez H. & Rios-Bolivar, M. (1994). Sliding Mode Control of dc-to-dc Power Converters via Extended Linearization, IEEE Trans Circ Syst-I, Vol. 41, No. 10 pp. 652-662. Sira-Ramirez, H.; Ortega, R.; Perez-Moreno, R.; Garcia-Esteban, M (1995). A sliding mode controller-observer for DC-to-DC power converters: a passivity approach. Proc. of IEEE Conf. Decision and Control, Vol. 4 Date: 13-15 Dec 1995, Pages: 3379 - 3384. Song T.-T. & Chung, H.S-H. (2008). Boundary control of boost converters using state-energy plane. IEEE Trans. Power Electr. Vol. 23, No. 2, pp. 551563 ISSN: 0885-8993. Sontag E.D. (1998). Mathematical Control Theory. Deterministic Finite Dimensional Systems. Springer Verlag, ISBN 0-387-98489-5 Sreekumar C. & Agarwal V. (2007). Hybrid control approach for the output voltage regulation in buck type DC-DC converter, IET Electr. Power Appl, vol 1, No. 6, pp 897-906. Sreekumar C. & Agarwal V. (2008). A hybrid control algorithm for voltage regulation in DC-DC Boost Converter, IEEE Trans. Industrial Electronics, vol 55, No. 6, pp 2530-2538. ISSN 0278-0046. Sreekumar C. &, Agarwal V. (2006). A Circuit Theoretical Approach to Hybrid Mode Switching Control of a Pseudo CCM Boost Converter, Proc IEEE International Conference on Industrial Technology, pp. 791-795, ISBN 1-4244-0726-5 , Numbai, India, 15-17 Dec. 2006 Sun, Z. & Ge, S.S. (2005). Switched Linear Systems. Control and design, Springer-Verlag, ISBN 1-852333-893-8, London. Hybrid Control of DC-DC Power Converters 195Renewable Energy196 Tan, S.C., Lai, Y.M. & Tse C.K. (2008). General design issues fo sliding mode controllers in DC-DC converters. IEEE Trans. Ind. Electronics. Vol. 55, No. 3, pp. 1160, 1174. ISSN: 0278-0046. Tao, H., Duarte, J.L. & Hendrix, M.A.M. (2006). Novel Zero-Voltage Switching Control Methods for a Multiple-Input Converter Interfacing a Fuel Cell and Supercapacitor. Proc. of Industrial Electronics Conference IEEE, No. 6-10, Nov. 2006 Page(s): 2341 – 2346. Ting-Ting Song & Chung, H.S.-h. (2008).Boundary Control of Boost Converters Using State-Energy Plane; IEEE Trans. Power Electronics, Vol. 23, No. 2, Page numbers (551 – 563), ISSN: 0885-8993. Tokat, S., Eksin, I., & Guzelkaya M. (2002). Sliding mode control using a nonlinear time-varying sliding surface. Proc. of the 10th Mediterranean Conference on Control and Automation - MED2002 Lisbon, Portugal, July 9-12, 2002. Tsinias, J (2007). Sufficient Conditions for Asymptotic Controllability and Hybrid Feedback Stabilization for a Class of Time-Varying Systems. IEEE Trans. Automatic Control, Vol. 52, No. 1, pp: 76 – 83, ISSN 0018-9286. Verghese,G.C., Kassakian, J.G. , Schlecht M-F. (1991) Principles of Power Electronics. Addison-Wesley, ISBN-10: 0201096897. NY. Xu, X. (2004). Practical stabilizability of a Class of Switched Systems. Proc. of American Control Conference , pp. 4537-4542, ISBN 0-7083-8335-4, Boston Massachuttes, July. Xu, X., He S. & Zhai, G. (2007). Practical uniform asymptotic stabilizability of switched systems with time-varying subsystems. Proc. of American Control Conference , pp. 681-686, ISBN 1-4244-0889-6, New York, July. Yang. T., (2001). Impulsive Control Theory. Lectures Notes in Control and Information Science. Springer Verlag ISBN: 3-540-422296-X Yang. T., (2001). Impulsive Systems and Control. Theory and Applications. Nova Science Publisher ISBN: 1-59033-058-7 You, Y.C. & Lee, E.B. (1989). Dynamical boundary control of two-dimensional wave equations: vibrating membrane on general domain. IEEE Trans. on Automatic Control, Vol. 34, No. 11, pp. 1181 – 1185, ISSN 0018-9286. Zhai, G. & Michel A.N. (2002) On practical of stability analysis of switched systems. Proc of IEEE Conf Decision and Control. pp. 3488-3493, ISBN 0-7803-7516-5, Vegas, Nevada. Zhai, G. & Michel A.N. (2003) Generalized practical stability analysis of discontinuous dynamical systems. Proc of IEEE Conf Decision and Control. pp. 1663-1668, ISBN 0-7803-7924-1, Maui, Hawaii. Interaction of Renewable Energy Source and Power Supply Network 197Interaction of Renewable Energy Source and Power Supply NetworkBranislav Dobrucký, Michal Pokorný and Mariana BeňováX Interaction of Renewable Energy Source and Power Supply Network Branislav Dobrucký, Michal Pokorný and Mariana Beňová University of Zilina Slovak Republic 1. Introduction to interaction of RES and network Renewable energy systems (RES) become more and more important as the serious sources, first of all from the point of view of improvement of power generating efficiency and effectiveness of operation. RES systems (in this sense) consist of renewable energy source, (e.g. photovoltaics, fuel cells, wind power,..), pre-conditioning unit, DC link, inverter, transformer (if necessary) and inductive coupling with electronic switch or direct connection to power supply network. The loads and supply system can be operated in three modes of operation: autonomous supply from the network or autonomous supply from RES system, and parallel operation of power supply network and RES(s). The chapter thereinafter deals with parallel operation of the both sources at steady-state and transient dynamic states. Legend: PSN – power supply network HV - high voltage LV - low voltage LD - loads ACC – energy accumulation BR - breakers ES - electronic switches L - separating inductors RES - renewable energy sources Fig. 1. Block diagram of power supply network and renewable energy system(s) connection LD1LD2ACCRES1ES1L1BR BR BR BRRES2L2LVHVPSNTRES211Renewable Energy196 Tan, S.C., Lai, Y.M. & Tse C.K. (2008). General design issues fo sliding mode controllers in DC-DC converters. IEEE Trans. Ind. Electronics. Vol. 55, No. 3, pp. 1160, 1174. ISSN: 0278-0046. Tao, H., Duarte, J.L. & Hendrix, M.A.M. (2006). Novel Zero-Voltage Switching Control Methods for a Multiple-Input Converter Interfacing a Fuel Cell and Supercapacitor. Proc. of Industrial Electronics Conference IEEE, No. 6-10, Nov. 2006 Page(s): 2341 – 2346. Ting-Ting Song & Chung, H.S.-h. (2008).Boundary Control of Boost Converters Using State-Energy Plane; IEEE Trans. Power Electronics, Vol. 23, No. 2, Page numbers (551 – 563), ISSN: 0885-8993. Tokat, S., Eksin, I., & Guzelkaya M. (2002). Sliding mode control using a nonlinear time-varying sliding surface. Proc. of the 10th Mediterranean Conference on Control and Automation - MED2002 Lisbon, Portugal, July 9-12, 2002. Tsinias, J (2007). Sufficient Conditions for Asymptotic Controllability and Hybrid Feedback Stabilization for a Class of Time-Varying Systems. IEEE Trans. Automatic Control, Vol. 52, No. 1, pp: 76 – 83, ISSN 0018-9286. Verghese,G.C., Kassakian, J.G. , Schlecht M-F. (1991) Principles of Power Electronics. Addison-Wesley, ISBN-10: 0201096897. NY. Xu, X. (2004). Practical stabilizability of a Class of Switched Systems. Proc. of American Control Conference , pp. 4537-4542, ISBN 0-7083-8335-4, Boston Massachuttes, July. Xu, X., He S. & Zhai, G. (2007). Practical uniform asymptotic stabilizability of switched systems with time-varying subsystems. Proc. of American Control Conference , pp. 681-686, ISBN 1-4244-0889-6, New York, July. Yang. T., (2001). Impulsive Control Theory. Lectures Notes in Control and Information Science. Springer Verlag ISBN: 3-540-422296-X Yang. T., (2001). Impulsive Systems and Control. Theory and Applications. Nova Science Publisher ISBN: 1-59033-058-7 You, Y.C. & Lee, E.B. (1989). Dynamical boundary control of two-dimensional wave equations: vibrating membrane on general domain. IEEE Trans. on Automatic Control, Vol. 34, No. 11, pp. 1181 – 1185, ISSN 0018-9286. Zhai, G. & Michel A.N. (2002) On practical of stability analysis of switched systems. Proc of IEEE Conf Decision and Control. pp. 3488-3493, ISBN 0-7803-7516-5, Vegas, Nevada. Zhai, G. & Michel A.N. (2003) Generalized practical stability analysis of discontinuous dynamical systems. Proc of IEEE Conf Decision and Control. pp. 1663-1668, ISBN 0-7803-7924-1, Maui, Hawaii. Interaction of Renewable Energy Source and Power Supply Network 197Interaction of Renewable Energy Source and Power Supply NetworkBranislav Dobrucký, Michal Pokorný and Mariana BeňováX Interaction of Renewable Energy Source and Power Supply Network Branislav Dobrucký, Michal Pokorný and Mariana Beňová University of Zilina Slovak Republic 1. Introduction to interaction of RES and network Renewable energy systems (RES) become more and more important as the serious sources, first of all from the point of view of improvement of power generating efficiency and effectiveness of operation. RES systems (in this sense) consist of renewable energy source, (e.g. photovoltaics, fuel cells, wind power,..), pre-conditioning unit, DC link, inverter, transformer (if necessary) and inductive coupling with electronic switch or direct connection to power supply network. The loads and supply system can be operated in three modes of operation: autonomous supply from the network or autonomous supply from RES system, and parallel operation of power supply network and RES(s). The chapter thereinafter deals with parallel operation of the both sources at steady-state and transient dynamic states. Legend: PSN – power supply network HV - high voltage LV - low voltage LD - loads ACC – energy accumulation BR - breakers ES - electronic switches L - separating inductors RES - renewable energy sources Fig. 1. Block diagram of power supply network and renewable energy system(s) connection LD1LD2ACCRES1ES1L1BR BR BR BRRES2L2LVHVPSNTRES211Renewable Energy198 Interconnection between renewable energy systems RES and power electric network depends on type and power of RES: - RES with nearly harmonic output voltage, synchronized by grid: direct connection, - RES with non-harmonic output voltage, synchronized by grid: inductive inter-connection, - remote RES non-synchronized with grid: HVDC energy transfer. Single-phase inverters are commonly used to obtain utility grade ac power in small distributed generation systems such as photovoltaics, wind power generators, fuel cells chip systems. When such single-phase systems are aggregated to form microgrids integrable with a three-phase inverter system or a common dc bus, it is desirable to maintain the aggregated power to be constant [Bala & Venkataramanan, 2007]. Various transfigurations of single-phase converter topologies brings [Xue et al, 2004]; [Koutroulis et al, 2001]; [Rajeshekara, 2005]. Large complex wind plants are not explicitly described in this chapter. Their behaviour and back influence on power supply network can be found in [CIGRE, 2007; Spacil, 2006]. The main characteristics of electric power quantities under sinusoidal, non-sinusoidal, balanced, or unbalanced conditions are described in the Standards [IEEE, 1999; IEEE, 2000]. Modelling and simulation of electric power quality parameters differs somewhat from the ordinary power system modelling as far as short -circuit, load flow and transient stability studies are concerned. The reason is that the behaviour of the system equipment must be predicted for frequencies well above the fundamental one [Dumitrescu, 2009; Ghartemani et al, 2004]. All above mentioned problems have to be taken in account during investigation of RES and power network, and proper methods should be used to fulfill the overall goals. 2. Theoretical background for single- and three-phase transients In general, transient phenomena of RES and power supply network can be investigated by different ways as follows: - as complex non-linear system, using high volume SW packages (PSCAD/EMTDC, EMTP, EDSA, OrCAD..) - as linearised system making possible of superposition: - in time domain (piece-wise linearization and state-variable method, z-transform and switching function method), - in frequency domain (decomposition into single harmonic linear subsystems) Simulation software packages are a powerful electromagnetic time domain transient simulation environment and study tools. Package usually includes static or dynamic characteristic of the electronic switches, and it works with invariant circuit topologic scheme and with parametric changes of the system during simulation. Above mentioned linearisation, superposition method and z-transformation can be conveniently used for investigated electric circuit’s analysis. Piece-wise linearisation in time domain. This method works during simulation with periodically variable circuit topologic structure of the system, and it uses electronic switches as separators between successive time intervals of operation of the system. Within a single time intervals of operation the topologic structure of the system is to be foreseen as constant one. Therefore one can use state variable method to describe RES system [Mohan et al, 2003; Dobrucky at al, 2007]: )()())((dd ttttuBxAx  (1) where: x(t) is the vector of state variables, A, B matrices of system elements, u(t) input vector of exciting functions  rii ttt0)( )()()( uDxCy (2) where: y(t) is the vector of output variables, C, D system matrices, r highest order of derivatives of the input vector (providing the derivatives exist). In the next time interval the state variables end-values of the previous time-interval will be considerate as the initial values. Note: Method of fictitious exciting function could be used in case of non-stationary elements of A matrix of the system. Using Park-Clarke orthogonal transform and subsequent two orthogonal Fourier series. Any m-phase system (symmetrical or non-symmetrical) can be transformed into equivalent 2-phase orthogonal system using Park-Clarke transform. Transform into stationary α, β- coordinate system is exclusively used in power electronics. That transform is defined by transformation equation  )(a)(a)(32)( 3221 tttt xxxx  (3) where  3π2jexp1a and real- and imaginary parts of function (3) for symmetrical system are  )()()(231)( 321α tttt xxxx  , (4)  )()(31)( 32β ttt xxx  . (5) Since x*(t) is complex time function (= x(t) +jx(t)) the time-waveforms of x(t) and x(t) can be expressed by complex Fourier series [Takeuchi, 1973; Bartsch, 1994]     1jj0j*ktkktkkktkk exexxextx (6) Properties of orthogonal Fourier series and convergence is described by [Marcokova, 1995; 2009]; and some application of them are given in [Takeuchi, 1973]; [Zaskalicky&Zaskalicka, 2008]. Using z-transform and switching function method. The renewables, as sources of pulse output voltage series, can be described by system of difference equations. For three-phase system the currents in α-β coordinates are given as )(g)(f)1( αT/6αT/6α nunini  (7) )(g)(f)1( T/6T/6 nunini   (8) Interaction of Renewable Energy Source and Power Supply Network 199Renewable Energy198 Interconnection between renewable energy systems RES and power electric network depends on type and power of RES: - RES with nearly harmonic output voltage, synchronized by grid: direct connection, - RES with non-harmonic output voltage, synchronized by grid: inductive inter-connection, - remote RES non-synchronized with grid: HVDC energy transfer. Single-phase inverters are commonly used to obtain utility grade ac power in small distributed generation systems such as photovoltaics, wind power generators, fuel cells chip systems. When such single-phase systems are aggregated to form microgrids integrable with a three-phase inverter system or a common dc bus, it is desirable to maintain the aggregated power to be constant [Bala & Venkataramanan, 2007]. Various transfigurations of single-phase converter topologies brings [Xue et al, 2004]; [Koutroulis et al, 2001]; [Rajeshekara, 2005]. Large complex wind plants are not explicitly described in this chapter. Their behaviour and back influence on power supply network can be found in [CIGRE, 2007; Spacil, 2006]. The main characteristics of electric power quantities under sinusoidal, non-sinusoidal, balanced, or unbalanced conditions are described in the Standards [IEEE, 1999; IEEE, 2000]. Modelling and simulation of electric power quality parameters differs somewhat from the ordinary power system modelling as far as short -circuit, load flow and transient stability studies are concerned. The reason is that the behaviour of the system equipment must be predicted for frequencies well above the fundamental one [Dumitrescu, 2009; Ghartemani et al, 2004]. All above mentioned problems have to be taken in account during investigation of RES and power network, and proper methods should be used to fulfill the overall goals. 2. Theoretical background for single- and three-phase transients In general, transient phenomena of RES and power supply network can be investigated by different ways as follows: - as complex non-linear system, using high volume SW packages (PSCAD/EMTDC, EMTP, EDSA, OrCAD..) - as linearised system making possible of superposition: - in time domain (piece-wise linearization and state-variable method, z-transform and switching function method), - in frequency domain (decomposition into single harmonic linear subsystems) Simulation software packages are a powerful electromagnetic time domain transient simulation environment and study tools. Package usually includes static or dynamic characteristic of the electronic switches, and it works with invariant circuit topologic scheme and with parametric changes of the system during simulation. Above mentioned linearisation, superposition method and z-transformation can be conveniently used for investigated electric circuit’s analysis. Piece-wise linearisation in time domain. This method works during simulation with periodically variable circuit topologic structure of the system, and it uses electronic switches as separators between successive time intervals of operation of the system. Within a single time intervals of operation the topologic structure of the system is to be foreseen as constant one. Therefore one can use state variable method to describe RES system [Mohan et al, 2003; Dobrucky at al, 2007]: )()())((dd ttttuBxAx  (1) where: x(t) is the vector of state variables, A, B matrices of system elements, u(t) input vector of exciting functions  rii ttt0)( )()()( uDxCy (2) where: y(t) is the vector of output variables, C, D system matrices, r highest order of derivatives of the input vector (providing the derivatives exist). In the next time interval the state variables end-values of the previous time-interval will be considerate as the initial values. Note: Method of fictitious exciting function could be used in case of non-stationary elements of A matrix of the system. Using Park-Clarke orthogonal transform and subsequent two orthogonal Fourier series. Any m-phase system (symmetrical or non-symmetrical) can be transformed into equivalent 2-phase orthogonal system using Park-Clarke transform. Transform into stationary α, β- coordinate system is exclusively used in power electronics. That transform is defined by transformation equation  )(a)(a)(32)( 3221 tttt xxxx  (3) where  3π2jexp1a and real- and imaginary parts of function (3) for symmetrical system are  )()()(231)( 321α tttt xxxx  , (4)  )()(31)( 32β ttt xxx  . (5) Since x*(t) is complex time function (= x(t) +jx(t)) the time-waveforms of x(t) and x(t) can be expressed by complex Fourier series [Takeuchi, 1973; Bartsch, 1994]     1jj0j*ktkktkkktkk exexxextx (6) Properties of orthogonal Fourier series and convergence is described by [Marcokova, 1995; 2009]; and some application of them are given in [Takeuchi, 1973]; [Zaskalicky&Zaskalicka, 2008]. Using z-transform and switching function method. The renewables, as sources of pulse output voltage series, can be described by system of difference equations. For three-phase system the currents in α-β coordinates are given as )(g)(f)1( αT/6αT/6α nunini  (7) )(g)(f)1( T/6T/6 nunini   (8) Interaction of Renewable Energy Source and Power Supply Network 199Renewable Energy200 where u(n) are the voltages in  and  coordinates, respectively Unnu  63sin32)(α (9) and Unnu  63cos32)(  (10) Waveform of exciting impulse function (= switching function) is shown in Figure 2 for 6- and 12-pulses voltages Fig. 2. Switching function of three-phase converter voltage with full width pulses (left)- and PWM modulation 12 pulses (right) The image of -component of 6-pulse voltage in z-plain is [Dobrucky et al, 2007]  11313)( 2323zzzzUzzzzUzU (11) Then, the image of -component of output current in z-plain is  )1()f(1g3)( 2T/6T/6 zzzzzUzI (12) Using inverse z-transform the discrete current series in time-domain will be obtained [Moravcik, 1992]. System decomposition into single harmonic linearised subsystems. Method of investi-gation assumes decomposition of real electric circuit into ν-harmonic separated equivalent schemes for each harmonic component [Dumitrescu, 2009]; [Benova, 2007]. Then transient analysis can be done for each scheme separately using 'impedance harmonic matrices', and each equivalent scheme is now linearised and therefore easily calculated. After finishing of calculation of each harmonic scheme, the effects of each investigated schemes are summed into resulting quantities of real non-linear electric circuit. That means that cumulative effect of sum of ν-harmonic circuits is superimposed on basic harmonic waveform with voltage source. The equivalent scheme for calculation of state quantities of the network with one appliance is drawn in Figure 3. Fig. 3. Equivalent scheme for calculation and state quantities of the network with one R-L appliance for all harmonic components, where ZT1(f) = f(νf1) and ZA1(f) = f(νf1) =R1+jν2πfL1, f1 is frequency of fundamental harmonic component, iv are current sources representing nonlinearity of appliance, R = ∑ Rv is the sum of source resistances This overall scheme is now decomposed into ν-separate schemes for each harmonic component. These schemes for fundamental- and ν- harmonic components of complex magnitudes will be as in Figure 4a and 4b: Fig. 4. a,b Equivalent scheme for fundamental (a) and ν - harmonic (b) components of one appliance, where Ū is complex magnitude of network voltage source with voltage of fundamental current source representing nonlinearity of appliance, Ī1 and Īν are complex magnitudes of network current for fundamental and ν- harmonic component, Īvν is complex magnitude of current source representing nonlinearity of appliance Complex magnitudes of ν-harmonic component can be obtained by harmonic analysis of its current using 'impedance (admittance) harmonic matrices' by nodal voltages for rated power of the appliance. For example, in case of ν - harmonic component of one appliance (equivalent scheme on Figure 3b) will be     v1A1A1T 2j112j1 IUπfLRRπfLRZ ii (13) where Ūν is nodal voltage (equal to appliance voltage in this case). Complex magnitude of network current for ν- harmonic component can be obtained using Li Ri i ZA1(f) iv3 u ~ ~ ivn ~ R ZT1(f) Li Ri Ī1 RA1 Ū ~ LA1 R1 ZT11 Li Ri Īν RA1 Īvν ~ LA1 Rν ZT1ν Interaction of Renewable Energy Source and Power Supply Network 201Renewable Energy200 where u(n) are the voltages in  and  coordinates, respectively Unnu  63sin32)(α (9) and Unnu  63cos32)(  (10) Waveform of exciting impulse function (= switching function) is shown in Figure 2 for 6- and 12-pulses voltages Fig. 2. Switching function of three-phase converter voltage with full width pulses (left)- and PWM modulation 12 pulses (right) The image of -component of 6-pulse voltage in z-plain is [Dobrucky et al, 2007]  11313)( 2323zzzzUzzzzUzU (11) Then, the image of -component of output current in z-plain is  )1()f(1g3)( 2T/6T/6 zzzzzUzI (12) Using inverse z-transform the discrete current series in time-domain will be obtained [Moravcik, 1992]. System decomposition into single harmonic linearised subsystems. Method of investi-gation assumes decomposition of real electric circuit into ν-harmonic separated equivalent schemes for each harmonic component [Dumitrescu, 2009]; [Benova, 2007]. Then transient analysis can be done for each scheme separately using 'impedance harmonic matrices', and each equivalent scheme is now linearised and therefore easily calculated. After finishing of calculation of each harmonic scheme, the effects of each investigated schemes are summed into resulting quantities of real non-linear electric circuit. That means that cumulative effect of sum of ν-harmonic circuits is superimposed on basic harmonic waveform with voltage source. The equivalent scheme for calculation of state quantities of the network with one appliance is drawn in Figure 3. Fig. 3. Equivalent scheme for calculation and state quantities of the network with one R-L appliance for all harmonic components, where ZT1(f) = f(νf1) and ZA1(f) = f(νf1) =R1+jν2πfL1, f1 is frequency of fundamental harmonic component, iv are current sources representing nonlinearity of appliance, R = ∑ Rv is the sum of source resistances This overall scheme is now decomposed into ν-separate schemes for each harmonic component. These schemes for fundamental- and ν- harmonic components of complex magnitudes will be as in Figure 4a and 4b: Fig. 4. a,b Equivalent scheme for fundamental (a) and ν - harmonic (b) components of one appliance, where Ū is complex magnitude of network voltage source with voltage of fundamental current source representing nonlinearity of appliance, Ī1 and Īν are complex magnitudes of network current for fundamental and ν- harmonic component, Īvν is complex magnitude of current source representing nonlinearity of appliance Complex magnitudes of ν-harmonic component can be obtained by harmonic analysis of its current using 'impedance (admittance) harmonic matrices' by nodal voltages for rated power of the appliance. For example, in case of ν - harmonic component of one appliance (equivalent scheme on Figure 3b) will be     v1A1A1T 2j112j1 IUπfLRRπfLRZ ii (13) where Ūν is nodal voltage (equal to appliance voltage in this case). Complex magnitude of network current for ν- harmonic component can be obtained using Li Ri i ZA1(f) iv3 u ~ ~ ivn ~ R ZT1(f) Li Ri Ī1 RA1 Ū ~ LA1 R1 ZT11 Li Ri Īν RA1 Īvν ~ LA1 Rν ZT1ν Interaction of Renewable Energy Source and Power Supply Network 201Renewable Energy202 ii πfLRZUI2j1T  (14) It is possible (after above decomposition) to calculate state quantities of each ν-harmonic. This overall scheme is now decomposed into ν-separate schemes for each harmonic component. Corresponding harmonic quantities from each scheme are synthesized (summarized) into final resulting waveform of investigated state quantity: itotal(t) = i1(t) + …….+ in (t) = in (t) (15) After synthesis of all harmonic components the total transient current is obtained. 3. Renewables and non-linear and linear passive and active loads 3.1 Harmonic output voltage RES with non-linear load during transients (island operation) Single-phase bridge rectifier supplied by stiff voltage (from network or RES with small inner impedance) has positive half-vawes on its DC side as depicted in Figure 5. Fig. 5. Scheme of single-phase RES with rectifying load and its DC side voltage DC voltage can be decomposed into Fourier series as [Bartsch, 1994]       ...4cos5312cos311214 ttUu m  (16) with DC component of mU2 . Supposing R-L load the corresponding current for single harmonic component can be calculated:.      tZUi cos , (17) where  22 LRZ   (18) RL arctan (19) Using synthesis of all harmonic components the total transient current is obtained U  t    /coscos tetZUi  (20) where τ = L / R. Note (nota bene): similarly for three-phase 6- pulse middle point rectifier       ...6cos7513cos4212133 ttUu m  (21) with DC component of 233 mU , and also for other types of rectifiers. Simulation results are given in Figure 6 and Figure 7 for transient and steady-state considering 999 harmonics. Fig. 6. Dynamic state of connecting nonlinear rectifying load to RES: voltage on DC side (blue) and DC current (green) Fig. 7. Steady-state time-vaweforms of DC voltage (blue) and current (green) 0 0.01 0.02 0.03 0.04 0.05 0.060 200400Time [s] Voltage [V]0 1020Current [A]0.1 0.102 0.104 0.106 0.108 0.11 0.112 0.114 0.116 0.118 0.120 200 400 Time [s]Voltage [V]. . 121416Current [A]Interaction of Renewable Energy Source and Power Supply Network 203Renewable Energy202 ii πfLRZUI2j1T  (14) It is possible (after above decomposition) to calculate state quantities of each ν-harmonic. This overall scheme is now decomposed into ν-separate schemes for each harmonic component. Corresponding harmonic quantities from each scheme are synthesized (summarized) into final resulting waveform of investigated state quantity: itotal(t) = i1(t) + …….+ in (t) = in (t) (15) After synthesis of all harmonic components the total transient current is obtained. 3. Renewables and non-linear and linear passive and active loads 3.1 Harmonic output voltage RES with non-linear load during transients (island operation) Single-phase bridge rectifier supplied by stiff voltage (from network or RES with small inner impedance) has positive half-vawes on its DC side as depicted in Figure 5. Fig. 5. Scheme of single-phase RES with rectifying load and its DC side voltage DC voltage can be decomposed into Fourier series as [Bartsch, 1994]       ...4cos5312cos311214 ttUu m  (16) with DC component of mU2 . Supposing R-L load the corresponding current for single harmonic component can be calculated:.      tZUi cos , (17) where  22 LRZ   (18) RL arctan (19) Using synthesis of all harmonic components the total transient current is obtained U  t    /coscos tetZUi  (20) where τ = L / R. Note (nota bene): similarly for three-phase 6- pulse middle point rectifier       ...6cos7513cos4212133 ttUu m  (21) with DC component of 233 mU , and also for other types of rectifiers. Simulation results are given in Figure 6 and Figure 7 for transient and steady-state considering 999 harmonics. Fig. 6. Dynamic state of connecting nonlinear rectifying load to RES: voltage on DC side (blue) and DC current (green) Fig. 7. Steady-state time-vaweforms of DC voltage (blue) and current (green) 0 0.01 0.02 0.03 0.04 0.05 0.060 200400Time [s] Voltage [V]0 1020Current [A]0.1 0.102 0.104 0.106 0.108 0.11 0.112 0.114 0.116 0.118 0.120 200 400 Time [s]Voltage [V]. . 121416Current [A]Interaction of Renewable Energy Source and Power Supply Network 203Renewable Energy204 3.2 Non-Harmonic output voltage RES with R-L load and active voltage source (parallel operation) Today’s converters provide very good quality output quantities regarding to their average and RMS values. Anyway, the output voltage is non-harmonic, switched by high frequency, Figure 8. U1L RAuL(t)uA0Ud2Ud20 Fig. 8. Scheme of single-phase converter in half-bridge connection (top), and its output voltage Simulation results for generating and regenerating regimes show Figure 9. 0 0.002 0.004 0.006 0.008 0.01 0.012 0.014 0.016 0.018 0.02-1-0.500.51Time [s] Uav0 [V]3*Iav0 [A]0 0.002 0.004 0.006 0.008 0.01 0.012 0.014 0.016 0.018 0.02-1-0.500.51Time [s] Uav03*Iav0 [A] Fig. 9. Time-waveforms of output voltage (blue) and current (red) for generating and regenerating regimes (considering 999 harmonics) 3.3 Three-phase non-harmonic output voltage RES with R-L load – behaviour prediction Three-phase 6-pulse inverter switching function is given in Figure 10. Fig. 10. Switching function for three-phase 6-pulse RES inverter Corresponding simulation results for complex and time domain are depicted in Figure 11. Fig. 11. Trajectories of output voltage in complex (left) and time domain (right) 3.4 Single-phase non-harmonic output voltage RES with R-L load – behaviour prediction Single-phase 2-pulse inverter switching function is given in Figure 12. Fig. 12. Switching function for single-phase 2-pulse RES inverter Corresponding simulation results for complex and time domain are depicted in Figure 13. -0,25-0,2-0,15-0,1-0,0500,050,10,150,20,25-0,3 -0,2 -0,1 0 0,10,20,30,4 -0,3000-0,2000-0,10000,00000,10000,20000,30000 6 12 18 24Interaction of Renewable Energy Source and Power Supply Network 205Renewable Energy204 3.2 Non-Harmonic output voltage RES with R-L load and active voltage source (parallel operation) Today’s converters provide very good quality output quantities regarding to their average and RMS values. Anyway, the output voltage is non-harmonic, switched by high frequency, Figure 8. U1L RAuL(t)uA0Ud2Ud20 Fig. 8. Scheme of single-phase converter in half-bridge connection (top), and its output voltage Simulation results for generating and regenerating regimes show Figure 9. 0 0.002 0.004 0.006 0.008 0.01 0.012 0.014 0.016 0.018 0.02-1-0.500.51Time [s] Uav0 [V]3*Iav0 [A]0 0.002 0.004 0.006 0.008 0.01 0.012 0.014 0.016 0.018 0.02-1-0.500.51Time [s] Uav03*Iav0 [A] Fig. 9. Time-waveforms of output voltage (blue) and current (red) for generating and regenerating regimes (considering 999 harmonics) 3.3 Three-phase non-harmonic output voltage RES with R-L load – behaviour prediction Three-phase 6-pulse inverter switching function is given in Figure 10. Fig. 10. Switching function for three-phase 6-pulse RES inverter Corresponding simulation results for complex and time domain are depicted in Figure 11. Fig. 11. Trajectories of output voltage in complex (left) and time domain (right) 3.4 Single-phase non-harmonic output voltage RES with R-L load – behaviour prediction Single-phase 2-pulse inverter switching function is given in Figure 12. Fig. 12. Switching function for single-phase 2-pulse RES inverter Corresponding simulation results for complex and time domain are depicted in Figure 13. -0,25-0,2-0,15-0,1-0,0500,050,10,150,20,25-0,3 -0,2 -0,1 0 0,10,20,30,4 -0,3000-0,2000-0,10000,00000,10000,20000,30000 6 12 18 24Interaction of Renewable Energy Source and Power Supply Network 205Renewable Energy206 Fig. 13. Trajectories of output voltage in complex (left) and time domain (right) 4. Example of a practical implementation of the introduced methods of solution for different types of operation of RES and power network The main task of RES system is to deliver power of different composition into power supply network: - active power delivery (as a priority mission), - and reactive or/and distortion powers delivery. Later possibility means that RES provides function of static compensator or/and power active filter. It depends on requests of power supply network what kind of power will be delivered by RES. The block schemes of connection of both sources and basic scheme of circuit configuration of single-phase voltage inverter are shown in Figure 14. Fig. 14. Basic connection of RES and power network, and circuit configuration of single-phase voltage inverter Simulation experiments have been carried out by MatLab simulation SW 2008b environ-ment using theory of power active filters [Singh et al, 1999; Dobrucky et al, 2006; Pavlanin et PSN source Compensating RES i res i load i source (non -linear) Load  al, 2008] and approach in subsection 3.2 above. Figure 15 shows simulation experiments of active compensation regime depending of time-instant of full load switching-on. Fig. 15. Simulation of RES in active compensation regime: Load switched-on at negative maximum of voltage (left) and in optimal instant of time (right) During compensation regime of RES the complementary compensating current ires is generated by such a way, that after its addition with non- sinusoidal and phase shifted load current iload, the only active power (e.g. active and harmonic current isource) will be delivered by supplying network: sourceload iii res  (22) The calculation of this compensating current is the most important activity of the active filter’s control circuit. The calculations can be carried out by different ways [Singh et al, 1999; Dobrucky et al, 2006; Pavlanin et al, 2008]. Simulation of RES in regime delivering of active power can be observed in subsection 3.2. Active power can be delivered to- or taken from power supply network. 5. The results of laboratory experiments There are oscilloscopic records from experimental verification of RES operation in Figs. 16, 17, 18, 19 and Figure 20 for both compensation and active power delivering. Fig. 16. Transient response of RES for the full load switched on: source current (top), load current (middle) and RES current (bottom) [Dobrucky et al, 2006] is uiloaifilter 0 2 4 6 8 1 0 2 4 6 8 1is uiloaifilter Interaction of Renewable Energy Source and Power Supply Network 207Renewable Energy206 Fig. 13. Trajectories of output voltage in complex (left) and time domain (right) 4. Example of a practical implementation of the introduced methods of solution for different types of operation of RES and power network The main task of RES system is to deliver power of different composition into power supply network: - active power delivery (as a priority mission), - and reactive or/and distortion powers delivery. Later possibility means that RES provides function of static compensator or/and power active filter. It depends on requests of power supply network what kind of power will be delivered by RES. The block schemes of connection of both sources and basic scheme of circuit configuration of single-phase voltage inverter are shown in Figure 14. Fig. 14. Basic connection of RES and power network, and circuit configuration of single-phase voltage inverter Simulation experiments have been carried out by MatLab simulation SW 2008b environ-ment using theory of power active filters [Singh et al, 1999; Dobrucky et al, 2006; Pavlanin et PSN source Compensating RES i res i load i source (non -linear) Load  al, 2008] and approach in subsection 3.2 above. Figure 15 shows simulation experiments of active compensation regime depending of time-instant of full load switching-on. Fig. 15. Simulation of RES in active compensation regime: Load switched-on at negative maximum of voltage (left) and in optimal instant of time (right) During compensation regime of RES the complementary compensating current ires is generated by such a way, that after its addition with non- sinusoidal and phase shifted load current iload, the only active power (e.g. active and harmonic current isource) will be delivered by supplying network: sourceload iii res  (22) The calculation of this compensating current is the most important activity of the active filter’s control circuit. The calculations can be carried out by different ways [Singh et al, 1999; Dobrucky et al, 2006; Pavlanin et al, 2008]. Simulation of RES in regime delivering of active power can be observed in subsection 3.2. Active power can be delivered to- or taken from power supply network. 5. The results of laboratory experiments There are oscilloscopic records from experimental verification of RES operation in Figs. 16, 17, 18, 19 and Figure 20 for both compensation and active power delivering. Fig. 16. Transient response of RES for the full load switched on: source current (top), load current (middle) and RES current (bottom) [Dobrucky et al, 2006] is uiloaifilter 0 2 4 6 8 1 0 2 4 6 8 1is uiloaifilter Interaction of Renewable Energy Source and Power Supply Network 207Renewable Energy208 Fig. 17. Switching-on of the RES delivering pure active current [bottom, red] into power supply network (top, yellow) Fig. 18. Steady-state operation of RES in regime of active power delivering: voltage (top) and current (bellow) Fig. 19. Transient record of RES in regime of power acceptor from network: network voltage (top, yellow), voltage of DC bus (top, green) of RES and current (bellow) Fig. 20. Steady-state operation of RES in regime of power acceptor from network: voltage (top) and current (bellow) 6. Evaluation and conclusion It has been shown that renewable energy source can work in two regimes of operation: as power generating or power consumption unit. Simulation and experimental results have proved excellent transient properties of the RES in both operations. As can be seen in Figure 16 and 17, respectively, the waveforms indicate an instantaneous reaction of compensation in generic regimes of RES. RES reacts already after the first calculation step t, so it is possible to use it as dynamic voltage restorer for small changes of the supply voltage. 7. References Bala, S. & Venkataramanan, G. (2007). Balanced Power Aggregation of Asymmetric Single-phase Systems. IEEE Power Electronics Specialists Conference (PESC’07), pp. 2772-2778, ISBN: 978-1-4244-0655-5, Orlando, FL, USA, June 2007 Bartsch, H.J. (1994). Mathematic Formulae, 4th revised edition (in German), Carl Hanser Publisher, ISBN 80-200-1448-9, Munich (DE) Benova, M. (2007): Non-sinusoidal Current and Voltage Analysis with regard to Power Condition of Electrical Machines (in Slovak), PhD thesis, Faculty of EE, University of Zilina (SK), pp. 115 Burger, B. & Engler, A. (2001). Fast signal conditioning in single phase systems, 9th European Conference on Power Electronics and Applications, pp. CD-ROM, ISBN, conference Graz (AT),1991 Dobrucky, B. Drozdy, S., Frivaldsky, M. & Spanik P. (2007). Interaction of Renewable Energy Source and Power Supply Network in Transient State, Proc. of ICCEP’07 Int’l Conf. on Clean Energy Power, pp. CD-ROM, Capri (IT), May 2007 Dobrucky, B., Marcokova, M., Pokorny, M. & Sul, R. (2007). Prediction Of Periodical Variable Structure System Behaviour Using Minimum Data Acquisition Time, Proc. of IASTED MIC’07 Int’l Conf. on Modelling, Identification, and Control, pp. CD-ROM, , Innsbruck (AT), Feb. 2007 Interaction of Renewable Energy Source and Power Supply Network 209Renewable Energy208 Fig. 17. Switching-on of the RES delivering pure active current [bottom, red] into power supply network (top, yellow) Fig. 18. Steady-state operation of RES in regime of active power delivering: voltage (top) and current (bellow) Fig. 19. Transient record of RES in regime of power acceptor from network: network voltage (top, yellow), voltage of DC bus (top, green) of RES and current (bellow) Fig. 20. Steady-state operation of RES in regime of power acceptor from network: voltage (top) and current (bellow) 6. Evaluation and conclusion It has been shown that renewable energy source can work in two regimes of operation: as power generating or power consumption unit. Simulation and experimental results have proved excellent transient properties of the RES in both operations. As can be seen in Figure 16 and 17, respectively, the waveforms indicate an instantaneous reaction of compensation in generic regimes of RES. RES reacts already after the first calculation step t, so it is possible to use it as dynamic voltage restorer for small changes of the supply voltage. 7. References Bala, S. & Venkataramanan, G. (2007). Balanced Power Aggregation of Asymmetric Single-phase Systems. IEEE Power Electronics Specialists Conference (PESC’07), pp. 2772-2778, ISBN: 978-1-4244-0655-5, Orlando, FL, USA, June 2007 Bartsch, H.J. (1994). Mathematic Formulae, 4th revised edition (in German), Carl Hanser Publisher, ISBN 80-200-1448-9, Munich (DE) Benova, M. (2007): Non-sinusoidal Current and Voltage Analysis with regard to Power Condition of Electrical Machines (in Slovak), PhD thesis, Faculty of EE, University of Zilina (SK), pp. 115 Burger, B. & Engler, A. (2001). Fast signal conditioning in single phase systems, 9th European Conference on Power Electronics and Applications, pp. CD-ROM, ISBN, conference Graz (AT),1991 Dobrucky, B. Drozdy, S., Frivaldsky, M. & Spanik P. (2007). Interaction of Renewable Energy Source and Power Supply Network in Transient State, Proc. of ICCEP’07 Int’l Conf. on Clean Energy Power, pp. CD-ROM, Capri (IT), May 2007 Dobrucky, B., Marcokova, M., Pokorny, M. & Sul, R. (2007). Prediction Of Periodical Variable Structure System Behaviour Using Minimum Data Acquisition Time, Proc. of IASTED MIC’07 Int’l Conf. on Modelling, Identification, and Control, pp. CD-ROM, , Innsbruck (AT), Feb. 2007 Interaction of Renewable Energy Source and Power Supply Network 209Renewable Energy210 Dobrucky, B. & Pokorny, M. (2006). Highest Dynamics and Ultra Fast Start-Up of Power Active Filter. International Review of Electrical Engineering, No. 8, Aug. 2006, pp. 391-399, ISSN 1827-6660 Dumitrescu, M.: Harmonic Digital Simulation Technique on an Industrial Environment. Electrical Review (Przeglad elektrotechniczny, PL), No. 1 (85), 2009, pp. 217-221, ISSN 0033-2097 Ghartemani, M. K., H. Mokhtari, M. R. Iravani, A Signal Processing System for Extraction of Harmonics and Reactive Current of Single-Phase Systems, IEEE Trans. on Power Delivery, Vol. 19, July 2004, pp. 979-986 IEEE, (1999). Standard 519-1992 Power Grids Distortion Limits. Voltage characteristics of electricity supplied by public distribution systems, European Norm EN 50160 IEEE (2000). Trial-Use Standard Definitions for the Measurement of Electric Power Quan-tities Under Sinusoidal, Nonsinusoidal, Balanced, or Unbalanced Conditions. IEEE Standard 1459 Koutroulis, E., K. Kalaitzakis, N.C. Voulgaris: Development of a Microcontroller-Based Photovoltaic Maximum power Point Tracking Control System. IEEE Trans, on Power Electronics, Vol. 16, No. 1, Jan. 2001, pp. 46-54 Marcokova, M. & Guldan V. (2009). On one orthogonal transform applied on a system of orthogonal polynomials in two variables. Journal of Applied Mathematics, Vol. 2, No. 2, pp. 239 - 245 Marcokova, M. (1995): Equiconvergence of Two Fourier Series. Journal of Approximation Theory, Vol. 80, No. 2, , pp. 151-16 Mohan, N., Undeland, T.M. & Robbins, W.P. (2003). Power Electronics : Converters, Applications, and Design 3rd Edition, John Wiley & Sons, Inc., ISBN 0-471-42908-2, New York Pavlanin, R. & Spanik, P. (2008). Effect of PCC Voltage Distortion on Shunt active Power filter Function (in Slovak), Proc. of 8th ERU08 Conference, on Electromagnetic Disturbation in Distributed- and Industrial Networks, pp. CD-ROM, ISBN 978-80-254-3821-3, Brno (CZ), Nov. 2008 Rajeshekara, K. (2005). Hybrid Fuel-Cell Strategies for Clean Power Generation. IEEE Trans, on Industry Application, Vol. 41, No. 3, May-June 2005, pp. 682-689, Singh, B., Al-Haddad, K., & Chandra, A. (1999). A review of active filters for power quality improvement. IEEE Transactions on Industrial Electronics, Vol. 46, pp. 960–971, Oct. 1999 Takeuchi, T.J. (1973). Theory of SCR Circuit and Application to Motor Control. Electrical Engineering College Press, 1968 - Russian Edition, Energia Publisher, Leningrad-St. Petersburg (RF), 248 pp., T 051(01)-73 130-73, Tokyo (JP) CIGRE Working Group (2007). C4.601 Modeling and Dynamic Behavior of Wind Generation Relates to Power System Control and Dynamic Performance, CIGRE, August 2007, Xue, Y., Chang, L., Kjaer, S. B., Bordonau, J. & Shimizu, T. (2004). Topologies of Single-phase Inverters for Small Distributed Power Generators: An Overview. IEEE Transactions on Power Electronics, Vol. 19, Sept. 2004, pp. 1305– 1314 Zaskalicky, P. & Zaskalicka M. (2009). Complex Fourier Series Model of a Three-Phase Voltage Source Inverter, Acta Electrotechnica et Informatica, Vol. N, No. X, 2009 (accepted) Marine Tidal Current Electric Power Generation: State of Art and Current Status 211Marine Tidal Current Electric Power Generation: State of Art and Current StatusYun Seng. Lim and Siong Lee. KohX Marine Tidal Current Electric Power Generation: State of Art and Current Status Yun Seng. Lim and Siong Lee. Koh University of Tunku Abdul Rahman Malaysia 1. Introduction Tidal power is a form of renewable energy generated by the periodic change in the ocean envelope of the Earth while interacting with Sun and the Moon via gravitational forces. The gravitational force of the Earth and the Moon attracts the ocean towards it. The motion of Earth around the centre of mass of the Earth-Moon system develops a bulge on the side of Earth opposite the Moon. The net effect of the two phenomena is two tidal bulges. The rotation of the Earth-Moon system cause the two bulges to rotate, making the sea level to rise and fall periodical to coastal observers. The tides created by the Earth-Moon system are known as lunar tides (Tom, 2007). The gravitational force of the Earth-Sun system has the same influence on the ocean. It creates bulges that tend to follow the Sun through the day. The influence of the Sun is about 46 % of that from the Moon. The positions of the solar bulges change much more slowly than the position of the lunar bulges. The lunar and solar tides will be additive, resulting in higher high tides and lower low tides. However, if the Moon, Earth and Sun form a right angle, the solar tide will tend to diminish the lunar tide, which is known as neap tides. This is because the moon’s contribution is more than twice that of the sun, the solar tide will not completely cancel the lunar tide. During neap tides, the high tides are not very high and low tides not very low. Neap tides occur at two week intervals. The large tides are caused by the linear alignment of the sun, Earth and moon which is also called spring tides. During spring tides, high tides are very high and low tides very low. Spring tides occur in two week intervals. Neap tides arrive a week after the spring tide. As Earth turns, landmasses obstruct the tidal crests, diverting, slowing and otherwise complicating their movements. This interference produces different patterns in the arrival of tidal crests at different places. The shape of the basin itself has a significant influence on the patterns and heights of tides. For these and other reasons, some coastlines experience semidiurnal tides: two high tides and two low tides of nearly equal level each day. Others have diurnal tides: one high and one low daily. The tidal pattern is called mixed if successive high tides or low tides are of significantly different heights throughout the cycle. This pattern is caused by blending diurnal and semidiurnal tides. 12Renewable Energy210 Dobrucky, B. & Pokorny, M. (2006). Highest Dynamics and Ultra Fast Start-Up of Power Active Filter. International Review of Electrical Engineering, No. 8, Aug. 2006, pp. 391-399, ISSN 1827-6660 Dumitrescu, M.: Harmonic Digital Simulation Technique on an Industrial Environment. Electrical Review (Przeglad elektrotechniczny, PL), No. 1 (85), 2009, pp. 217-221, ISSN 0033-2097 Ghartemani, M. K., H. Mokhtari, M. R. Iravani, A Signal Processing System for Extraction of Harmonics and Reactive Current of Single-Phase Systems, IEEE Trans. on Power Delivery, Vol. 19, July 2004, pp. 979-986 IEEE, (1999). Standard 519-1992 Power Grids Distortion Limits. Voltage characteristics of electricity supplied by public distribution systems, European Norm EN 50160 IEEE (2000). Trial-Use Standard Definitions for the Measurement of Electric Power Quan-tities Under Sinusoidal, Nonsinusoidal, Balanced, or Unbalanced Conditions. IEEE Standard 1459 Koutroulis, E., K. Kalaitzakis, N.C. Voulgaris: Development of a Microcontroller-Based Photovoltaic Maximum power Point Tracking Control System. IEEE Trans, on Power Electronics, Vol. 16, No. 1, Jan. 2001, pp. 46-54 Marcokova, M. & Guldan V. (2009). On one orthogonal transform applied on a system of orthogonal polynomials in two variables. Journal of Applied Mathematics, Vol. 2, No. 2, pp. 239 - 245 Marcokova, M. (1995): Equiconvergence of Two Fourier Series. Journal of Approximation Theory, Vol. 80, No. 2, , pp. 151-16 Mohan, N., Undeland, T.M. & Robbins, W.P. (2003). Power Electronics : Converters, Applications, and Design 3rd Edition, John Wiley & Sons, Inc., ISBN 0-471-42908-2, New York Pavlanin, R. & Spanik, P. (2008). Effect of PCC Voltage Distortion on Shunt active Power filter Function (in Slovak), Proc. of 8th ERU08 Conference, on Electromagnetic Disturbation in Distributed- and Industrial Networks, pp. CD-ROM, ISBN 978-80-254-3821-3, Brno (CZ), Nov. 2008 Rajeshekara, K. (2005). Hybrid Fuel-Cell Strategies for Clean Power Generation. IEEE Trans, on Industry Application, Vol. 41, No. 3, May-June 2005, pp. 682-689, Singh, B., Al-Haddad, K., & Chandra, A. (1999). A review of active filters for power quality improvement. IEEE Transactions on Industrial Electronics, Vol. 46, pp. 960–971, Oct. 1999 Takeuchi, T.J. (1973). Theory of SCR Circuit and Application to Motor Control. Electrical Engineering College Press, 1968 - Russian Edition, Energia Publisher, Leningrad-St. Petersburg (RF), 248 pp., T 051(01)-73 130-73, Tokyo (JP) CIGRE Working Group (2007). C4.601 Modeling and Dynamic Behavior of Wind Generation Relates to Power System Control and Dynamic Performance, CIGRE, August 2007, Xue, Y., Chang, L., Kjaer, S. B., Bordonau, J. & Shimizu, T. (2004). Topologies of Single-phase Inverters for Small Distributed Power Generators: An Overview. IEEE Transactions on Power Electronics, Vol. 19, Sept. 2004, pp. 1305– 1314 Zaskalicky, P. & Zaskalicka M. (2009). Complex Fourier Series Model of a Three-Phase Voltage Source Inverter, Acta Electrotechnica et Informatica, Vol. N, No. X, 2009 (accepted) Marine Tidal Current Electric Power Generation: State of Art and Current Status 211Marine Tidal Current Electric Power Generation: State of Art and Current StatusYun Seng. Lim and Siong Lee. KohX Marine Tidal Current Electric Power Generation: State of Art and Current Status Yun Seng. Lim and Siong Lee. Koh University of Tunku Abdul Rahman Malaysia 1. Introduction Tidal power is a form of renewable energy generated by the periodic change in the ocean envelope of the Earth while interacting with Sun and the Moon via gravitational forces. The gravitational force of the Earth and the Moon attracts the ocean towards it. The motion of Earth around the centre of mass of the Earth-Moon system develops a bulge on the side of Earth opposite the Moon. The net effect of the two phenomena is two tidal bulges. The rotation of the Earth-Moon system cause the two bulges to rotate, making the sea level to rise and fall periodical to coastal observers. The tides created by the Earth-Moon system are known as lunar tides (Tom, 2007). The gravitational force of the Earth-Sun system has the same influence on the ocean. It creates bulges that tend to follow the Sun through the day. The influence of the Sun is about 46 % of that from the Moon. The positions of the solar bulges change much more slowly than the position of the lunar bulges. The lunar and solar tides will be additive, resulting in higher high tides and lower low tides. However, if the Moon, Earth and Sun form a right angle, the solar tide will tend to diminish the lunar tide, which is known as neap tides. This is because the moon’s contribution is more than twice that of the sun, the solar tide will not completely cancel the lunar tide. During neap tides, the high tides are not very high and low tides not very low. Neap tides occur at two week intervals. The large tides are caused by the linear alignment of the sun, Earth and moon which is also called spring tides. During spring tides, high tides are very high and low tides very low. Spring tides occur in two week intervals. Neap tides arrive a week after the spring tide. As Earth turns, landmasses obstruct the tidal crests, diverting, slowing and otherwise complicating their movements. This interference produces different patterns in the arrival of tidal crests at different places. The shape of the basin itself has a significant influence on the patterns and heights of tides. For these and other reasons, some coastlines experience semidiurnal tides: two high tides and two low tides of nearly equal level each day. Others have diurnal tides: one high and one low daily. The tidal pattern is called mixed if successive high tides or low tides are of significantly different heights throughout the cycle. This pattern is caused by blending diurnal and semidiurnal tides. 12Renewable Energy212 Fig. 1. Lunar tide created by the Earth-Moon System Humans have found ways to use the tides. As early as the eight century, the Spanish, French and British built tidal storage ponds behind dams that were filled by the incoming tidal through sluice gates. These gates were closed at high tide and the trapped water then released to the sea through a water wheel to mill grain. The Eling Tide Mill in the United Kingdom has been producing flour with tidal power since 1418 (Eling, 2009). A few tidal power electrical generation plants have been built that operate on the similar principle, in that a dam or barrage is constructed to impound the water at high tide and then releasing the water through openings with conventional hydroelectric turbine-generators when the water in the dam has sufficient potential energy. Although tidal power has not been widely used, tidal power is more reliable and predictable than wind energy and solar power. The Earth's tides are caused by the tidal forces due to gravitational interaction with the Moon and Sun, and the Earth's rotation, tidal power is practically inexhaustible and classified as a renewable energy source. Numerous techniques have been proposed to extract energy from tidal current energy. The kinetic energy present in marine and tidal currents can be converted to electricity using relatively conventional turbine technology. There are several sites of the global ocean have the potential tidal energy that can be harnessed economically. Some of those sites are Pembrokeshire in Wales, River Severn between Wales and England, Cook Strait in New Zealand, Bay of Fundy in Canada, East River in New York City, Golden Gate in the San Francisco Bay, Piscataqua River in New Hampshire, and The Race of Alderney and The Swinge in the Channel Islands. Those sites can make a significant contribution to electricity supply. This is the reason why the marine renewable sector is currently the focus of much industrial and academic research around the world. There are generally three categories of tidal energy technologies as described in the following. 2. Categories of tidal energy technologies Tidal power can be classified into three main types: Barrage tidal power: A physical barrier, namely the Barrage, is constructed within the sea with Sluice Gates to control the flow of sea water. The Sluice Gates are to be closed at high tide so that the water level inside the barrage is held at its highest level. As the tide recedes, a difference in water level in between the barrage and the sea is created. The potential energy from the water level difference can then drive turbines to generate electricity. Fig. 2. Barrage tidal power for electricity generation (Source: http://home.clara.net/ darvill/altenerg/tidal.htm) Tidal stream system: A horizontal axis turbines are placed in the path of tidal currents to generate electricity, similar to the operation of wind turbine. This method is getting popular because the cost and ecological impact of tidal turbines is much lower than that of the barrage system. Marine Tidal Current Electric Power Generation: State of Art and Current Status 213Renewable Energy212 Fig. 1. Lunar tide created by the Earth-Moon System Humans have found ways to use the tides. As early as the eight century, the Spanish, French and British built tidal storage ponds behind dams that were filled by the incoming tidal through sluice gates. These gates were closed at high tide and the trapped water then released to the sea through a water wheel to mill grain. The Eling Tide Mill in the United Kingdom has been producing flour with tidal power since 1418 (Eling, 2009). A few tidal power electrical generation plants have been built that operate on the similar principle, in that a dam or barrage is constructed to impound the water at high tide and then releasing the water through openings with conventional hydroelectric turbine-generators when the water in the dam has sufficient potential energy. Although tidal power has not been widely used, tidal power is more reliable and predictable than wind energy and solar power. The Earth's tides are caused by the tidal forces due to gravitational interaction with the Moon and Sun, and the Earth's rotation, tidal power is practically inexhaustible and classified as a renewable energy source. Numerous techniques have been proposed to extract energy from tidal current energy. The kinetic energy present in marine and tidal currents can be converted to electricity using relatively conventional turbine technology. There are several sites of the global ocean have the potential tidal energy that can be harnessed economically. Some of those sites are Pembrokeshire in Wales, River Severn between Wales and England, Cook Strait in New Zealand, Bay of Fundy in Canada, East River in New York City, Golden Gate in the San Francisco Bay, Piscataqua River in New Hampshire, and The Race of Alderney and The Swinge in the Channel Islands. Those sites can make a significant contribution to electricity supply. This is the reason why the marine renewable sector is currently the focus of much industrial and academic research around the world. There are generally three categories of tidal energy technologies as described in the following. 2. Categories of tidal energy technologies Tidal power can be classified into three main types: Barrage tidal power: A physical barrier, namely the Barrage, is constructed within the sea with Sluice Gates to control the flow of sea water. The Sluice Gates are to be closed at high tide so that the water level inside the barrage is held at its highest level. As the tide recedes, a difference in water level in between the barrage and the sea is created. The potential energy from the water level difference can then drive turbines to generate electricity. Fig. 2. Barrage tidal power for electricity generation (Source: http://home.clara.net/ darvill/altenerg/tidal.htm) Tidal stream system: A horizontal axis turbines are placed in the path of tidal currents to generate electricity, similar to the operation of wind turbine. This method is getting popular because the cost and ecological impact of tidal turbines is much lower than that of the barrage system. Marine Tidal Current Electric Power Generation: State of Art and Current Status 213Renewable Energy214 Fig. 3. Tidal stream turbine (Source: http://www.altdotenergy.com/2008/12/seagen-achieves-maximum-capacity-of-12-mw/) Tidal lagoons: These approaches are similar to barrages, but can be constructed as self contained structures, not fully across an estuary, and are claimed to incur much lower cost and impact overall. Furthermore they can be configured to generate electricity continuously which is not the case with barrages. Fig. 4. Tidal lagoon with 3 pools configuration (Source: http://www.tidalelectric.com/ technology.html) 3. Barrage tidal power There are only a few barrage tidal power plants operating globally. The largest one is 240 MW plant on the Rance River in France, and two small plants on the Bay of Fundy and across a tiny inlet in Kislaya Guba Russia. The Rance tidal power plant has been operating since 1966 with the annual production of 600 GWh (Charlie, 2007). The barrage system consists of caissons, embankments, sluices, turbines, and ship locks. Caisson is a very large concrete blocks used to house the sluices, turbines, and ship locks. Embankments seal a basin where it is not sealed by caissons. The sluice gates used in tidal power are the flap gate, vertical rising gate, radial gate, and rising sector. Fig. 5. Barrage tidal power plant on Rance River (Source: http://www.britannica.com/ EBchecked/topic-art/595132/118418/Tidal-power-generation-station-on-the-Rance-River-in-Saint) The barrage system generates electricity using the difference in water level in between the barrage and the sea. The difference in water level is created by a process known as outflow or ebb generation where the basin is filled through the sluice gates during high tide while the turbine gates are closed as shown in Figure 6. Then the sluice gates are closed when a significant difference in water level is achieved. The turbine gates are opened to let the water to flow out, hence generating electricity. This cycle repeats itself. If the difference in water level is not great enough, pumping may operate where turbines are powered by the grid electricity to pump water from the sea into the basin in order to raise the height by, say, another 2 ft (61 cm). The cost for creating the extra 2 ft rise can be returned by the total electricity generation as a result of the pumping. Fig. 6. Ebb generation scheme Marine Tidal Current Electric Power Generation: State of Art and Current Status 215Renewable Energy214 Fig. 3. Tidal stream turbine (Source: http://www.altdotenergy.com/2008/12/seagen-achieves-maximum-capacity-of-12-mw/) Tidal lagoons: These approaches are similar to barrages, but can be constructed as self contained structures, not fully across an estuary, and are claimed to incur much lower cost and impact overall. Furthermore they can be configured to generate electricity continuously which is not the case with barrages. Fig. 4. Tidal lagoon with 3 pools configuration (Source: http://www.tidalelectric.com/ technology.html) 3. Barrage tidal power There are only a few barrage tidal power plants operating globally. The largest one is 240 MW plant on the Rance River in France, and two small plants on the Bay of Fundy and across a tiny inlet in Kislaya Guba Russia. The Rance tidal power plant has been operating since 1966 with the annual production of 600 GWh (Charlie, 2007). The barrage system consists of caissons, embankments, sluices, turbines, and ship locks. Caisson is a very large concrete blocks used to house the sluices, turbines, and ship locks. Embankments seal a basin where it is not sealed by caissons. The sluice gates used in tidal power are the flap gate, vertical rising gate, radial gate, and rising sector. Fig. 5. Barrage tidal power plant on Rance River (Source: http://www.britannica.com/ EBchecked/topic-art/595132/118418/Tidal-power-generation-station-on-the-Rance-River-in-Saint) The barrage system generates electricity using the difference in water level in between the barrage and the sea. The difference in water level is created by a process known as outflow or ebb generation where the basin is filled through the sluice gates during high tide while the turbine gates are closed as shown in Figure 6. Then the sluice gates are closed when a significant difference in water level is achieved. The turbine gates are opened to let the water to flow out, hence generating electricity. This cycle repeats itself. If the difference in water level is not great enough, pumping may operate where turbines are powered by the grid electricity to pump water from the sea into the basin in order to raise the height by, say, another 2 ft (61 cm). The cost for creating the extra 2 ft rise can be returned by the total electricity generation as a result of the pumping. Fig. 6. Ebb generation scheme Marine Tidal Current Electric Power Generation: State of Art and Current Status 215Renewable Energy216 Electricity can also be generated through inflow generation where the basin is filled through the turbines instead of sluice gates during high tide. However, this approach is generally less efficient than outflow generation, because the volume contained in the basin is less than the volume contained through outflow generation. Therefore the amount of electricity generated using this method is less than that of outflow generation. 3.1 Environmental impact Barrage systems have not been widely used because it may create several environmental issues. The placement of a barrage into an estuary alters the flow of saltwater in and out of estuary which eventually change the hydrology and salinity, hence creating negative effects on the marine mammals that use the estuaries as their habitats (Pelc et al., 2002). During the construction of the tidal barrage, the estuary was isolated from the sea water, hence damaging flora and fauna. Some species lost their habitat due to La Rance’s construction. Also as a result of the construction, sandbanks disappeared, the beach of St. Servan was badly damaged. (Charlie, 2007). Estuaries often have high volume of sediments moving through them, from the rivers to the sea. The introduction of a barrage into an estuary may result in sediment accumulation within the barrage, affecting the ecosystem and also the operation of the barrage. Another environmental impact of barrage system is fish kill. Fish may move through sluices safely, but when these are closed, fish will seek out turbines and attempt to swim through them. If the turbines are moving slowly enough, such as low velocities of 25-50 rpm, fish kill can be minimised. However, some fish will be unable to escape the water speed near a turbine and will be sucked through. Fish can be killed by pressure drop, contact with blades, cavitation, etc. The most fish-friendly turbine design has fish mortality per pass of approximately 15%. Various passage technologies (fish ladders, fish lifts, etc.) have so far failed to overcome this problem for tidal barrages. Research in sonic guidance of fish is ongoing (Peltier, 2003). The Open-Centre turbine may minimise this problem by allowing fish to pass through the open centre of the turbine. Fig. 7. Open centre tidal turbine (Source: http://www.greentechmedia.com/articles/ photos-clean-power-at-the-edge-of-the-world-5723.html) 3.2 Economics Tidal barrage power schemes have a high capital cost and a very low running cost. Governments may be able to finance tidal barrage power, but many are not willing to do so because they need to wait for many years before investment return and may face high irreversible environmental and social issues. For example, the energy policy of the United Kingdom (European Commission, 2008) realises the potential role of tidal energy and expresses the need for local councils to understand the possible contribution of tidal projects to meeting the national goal of renewable energy. The UK government itself appreciates the technical viability, but has failed to provide meaningful incentives to move these goals forward. 4. Tidal stream generators Tidal stream generators draw energy from currents, which is the same way as wind turbines. The higher density of water, 832 times the density of air, means that a single generator can provide significant power at low tidal flow velocities as compared with wind speed. In another words, water speeds of nearly one-tenth of the speed of wind provide the same power for the same size of turbine system. The total kinetic power in a marine current can be expressed by the following equation. 3Av21P  (1) Where ρ is the seawater density (kgm-3), A is the swept area of turbine blades (m2) and v is the velocity of seawater (m/s). However, a marine energy converter or turbine can only harness a fraction of this power due to losses and Eq. (1) is modified as follows. 3pAvC21P  Where Cp is the power coefficient and the percentage of power that can be extracted from the fluid stream taking into account losses due to Betz’s law and the internal mechanisms within the converter or turbine. For wind generators, Cp has typical values in the range 0.25-0.3. The upper limit is for highly efficient machines with low mechanical losses. For marine turbines, Cp is in the range of 0.35-0.5. Tidal stream generators are not fully mature because, at present, there is no commercial scale production of tidal generators and no standard technology being recognised as the clear winner. A variety of designs are being experimented. Several prototypes have been developed. Some are very close to large scale deployment. However, their efficiencies and feasibility have yet to be independently verified. At present, there are several prototypes, namely horizontal axis turbine, vertical axis turbine, oscillating devices and tidal turbine with venturi shroud as described in the following. 4.1 Horizontal axis turbines These horizontal axis turbines have the similar concept to traditional windmills operating under the sea and have the most prototypes currently operating. A prototype of a 300 kW horizontal axis turbine was installed in Kvalsund, south of Hammerfest, Norway (Penman, 2009) and connected to the grid on 13 November 2003. Marine Tidal Current Electric Power Generation: State of Art and Current Status 217Renewable Energy216 Electricity can also be generated through inflow generation where the basin is filled through the turbines instead of sluice gates during high tide. However, this approach is generally less efficient than outflow generation, because the volume contained in the basin is less than the volume contained through outflow generation. Therefore the amount of electricity generated using this method is less than that of outflow generation. 3.1 Environmental impact Barrage systems have not been widely used because it may create several environmental issues. The placement of a barrage into an estuary alters the flow of saltwater in and out of estuary which eventually change the hydrology and salinity, hence creating negative effects on the marine mammals that use the estuaries as their habitats (Pelc et al., 2002). During the construction of the tidal barrage, the estuary was isolated from the sea water, hence damaging flora and fauna. Some species lost their habitat due to La Rance’s construction. Also as a result of the construction, sandbanks disappeared, the beach of St. Servan was badly damaged. (Charlie, 2007). Estuaries often have high volume of sediments moving through them, from the rivers to the sea. The introduction of a barrage into an estuary may result in sediment accumulation within the barrage, affecting the ecosystem and also the operation of the barrage. Another environmental impact of barrage system is fish kill. Fish may move through sluices safely, but when these are closed, fish will seek out turbines and attempt to swim through them. If the turbines are moving slowly enough, such as low velocities of 25-50 rpm, fish kill can be minimised. However, some fish will be unable to escape the water speed near a turbine and will be sucked through. Fish can be killed by pressure drop, contact with blades, cavitation, etc. The most fish-friendly turbine design has fish mortality per pass of approximately 15%. Various passage technologies (fish ladders, fish lifts, etc.) have so far failed to overcome this problem for tidal barrages. Research in sonic guidance of fish is ongoing (Peltier, 2003). The Open-Centre turbine may minimise this problem by allowing fish to pass through the open centre of the turbine. Fig. 7. Open centre tidal turbine (Source: http://www.greentechmedia.com/articles/ photos-clean-power-at-the-edge-of-the-world-5723.html) 3.2 Economics Tidal barrage power schemes have a high capital cost and a very low running cost. Governments may be able to finance tidal barrage power, but many are not willing to do so because they need to wait for many years before investment return and may face high irreversible environmental and social issues. For example, the energy policy of the United Kingdom (European Commission, 2008) realises the potential role of tidal energy and expresses the need for local councils to understand the possible contribution of tidal projects to meeting the national goal of renewable energy. The UK government itself appreciates the technical viability, but has failed to provide meaningful incentives to move these goals forward. 4. Tidal stream generators Tidal stream generators draw energy from currents, which is the same way as wind turbines. The higher density of water, 832 times the density of air, means that a single generator can provide significant power at low tidal flow velocities as compared with wind speed. In another words, water speeds of nearly one-tenth of the speed of wind provide the same power for the same size of turbine system. The total kinetic power in a marine current can be expressed by the following equation. 3Av21P  (1) Where ρ is the seawater density (kgm-3), A is the swept area of turbine blades (m2) and v is the velocity of seawater (m/s). However, a marine energy converter or turbine can only harness a fraction of this power due to losses and Eq. (1) is modified as follows. 3pAvC21P  Where Cp is the power coefficient and the percentage of power that can be extracted from the fluid stream taking into account losses due to Betz’s law and the internal mechanisms within the converter or turbine. For wind generators, Cp has typical values in the range 0.25-0.3. The upper limit is for highly efficient machines with low mechanical losses. For marine turbines, Cp is in the range of 0.35-0.5. Tidal stream generators are not fully mature because, at present, there is no commercial scale production of tidal generators and no standard technology being recognised as the clear winner. A variety of designs are being experimented. Several prototypes have been developed. Some are very close to large scale deployment. However, their efficiencies and feasibility have yet to be independently verified. At present, there are several prototypes, namely horizontal axis turbine, vertical axis turbine, oscillating devices and tidal turbine with venturi shroud as described in the following. 4.1 Horizontal axis turbines These horizontal axis turbines have the similar concept to traditional windmills operating under the sea and have the most prototypes currently operating. A prototype of a 300 kW horizontal axis turbine was installed in Kvalsund, south of Hammerfest, Norway (Penman, 2009) and connected to the grid on 13 November 2003. Marine Tidal Current Electric Power Generation: State of Art and Current Status 217Renewable Energy218 Fig. 8. Tidal turbine in comparison with offshore wind turbine A 300 kW horizontal axis turbine, also known as Seaflow, was installed by Marine Current Turbines off the coast of Lynmouth, Devon, England, in 2003 (Fraenkel, 2004). The turbine has a diameter of 11m being fitted to a steel pile which was driven into the seabed. It was connected to a dump load. Fig. 9. Seaflow Turbine A prototype project was installed in the East River between Queens and Roosevelt Island in New York City in the United States in April 2007 (Verdant Power, 2009). However, the blades broke off due to the strong tidal currents. Therefore, new reinforced turbines were installed in September 2008. (Urbina, 2004) (Galbraith, 2008). A prototype, called SeaGen, was installed by Marine Current Turbines in Strangford Lough in Northern Ireland in April 2008. The turbine began to generate at full power of about 1.2 MW in December 2008 and was reported to feed 150kW into the grid for the first time on 17 July 2008. It is currently the only commercial scale device to have been installed anywhere in the world (MCT, 2008). Fig. 10. Seagen Farm A prototype has been developed by OpenHydro, an Irish company (The Renewable Energy Centre, 2008). Its performance is being studied at the European Marine Energy Centre (EMEC), in Orkney, Scotland. Fig. 11. Seagen installed by Marine Current Turbine in Strangford Lough in Northern Ireland (Source: http://www.iconocast.com/00001/G1/News4.htm) 4.2 Vertical axis turbines The Gorlov helical turbine (GHT), evolved from the Darrieus turbine design by altering its blade to become helical blades. This cross-flow turbine was developed in 1994. The turbine consists of one or more long helical blades that run along a cylindrical surface like a screw thread that has ‘airplane wing’ profile. The blades provide a reaction thrust that can rotate the turbine. The turbine shaft (axis of rotation) should be perpendicular to the water current, and the turbine can be positioned either horizontally or vertically. Due to its axial symmetry, the turbine always develops unidirectional rotation, even in reversible tidal currents. This project has been proposed for the Uldolmok Strait in Korea, where a very strong reversible tidal current up to 6ms-1 (Cascio, 2005). Marine Tidal Current Electric Power Generation: State of Art and Current Status 219Renewable Energy218 Fig. 8. Tidal turbine in comparison with offshore wind turbine A 300 kW horizontal axis turbine, also known as Seaflow, was installed by Marine Current Turbines off the coast of Lynmouth, Devon, England, in 2003 (Fraenkel, 2004). The turbine has a diameter of 11m being fitted to a steel pile which was driven into the seabed. It was connected to a dump load. Fig. 9. Seaflow Turbine A prototype project was installed in the East River between Queens and Roosevelt Island in New York City in the United States in April 2007 (Verdant Power, 2009). However, the blades broke off due to the strong tidal currents. Therefore, new reinforced turbines were installed in September 2008. (Urbina, 2004) (Galbraith, 2008). A prototype, called SeaGen, was installed by Marine Current Turbines in Strangford Lough in Northern Ireland in April 2008. The turbine began to generate at full power of about 1.2 MW in December 2008 and was reported to feed 150kW into the grid for the first time on 17 July 2008. It is currently the only commercial scale device to have been installed anywhere in the world (MCT, 2008). Fig. 10. Seagen Farm A prototype has been developed by OpenHydro, an Irish company (The Renewable Energy Centre, 2008). Its performance is being studied at the European Marine Energy Centre (EMEC), in Orkney, Scotland. Fig. 11. Seagen installed by Marine Current Turbine in Strangford Lough in Northern Ireland (Source: http://www.iconocast.com/00001/G1/News4.htm) 4.2 Vertical axis turbines The Gorlov helical turbine (GHT), evolved from the Darrieus turbine design by altering its blade to become helical blades. This cross-flow turbine was developed in 1994. The turbine consists of one or more long helical blades that run along a cylindrical surface like a screw thread that has ‘airplane wing’ profile. The blades provide a reaction thrust that can rotate the turbine. The turbine shaft (axis of rotation) should be perpendicular to the water current, and the turbine can be positioned either horizontally or vertically. Due to its axial symmetry, the turbine always develops unidirectional rotation, even in reversible tidal currents. This project has been proposed for the Uldolmok Strait in Korea, where a very strong reversible tidal current up to 6ms-1 (Cascio, 2005). Marine Tidal Current Electric Power Generation: State of Art and Current Status 219Renewable Energy220 Fig. 12. Gorlov turbine (Source: http://www.oce.uri.edu/oce311/) Neptune Renewable Energy has developed Proteus (Neptune, 2009). The Neptune Proteus is a vertical axis turbine mounted within a symmetrical Venturi diffuser duct and beneath a very simple steel deck and buoyancy chambers. The Neptune Proteus is designed for estuarine sites that have high tidal currents, yet have the advantages of lower access, cabling and maintenance costs than in offshore environments. The vertical shaft connects to the 1:200 gearbox and generator in the deck housing. The device is moored in the free stream, minimising environmental impact and operates equally efficiently for both flood and ebb currents. The rotor is maintained at optimal power outputs by sets of computer controlled shutters within the duct and by the variable electrical load. The overall efficiency of the system suggests to be greater than 45%. Fig. 13. The Neptune Proteus The Enermar Project developed Kobold turbine (Ponte, 2009) that can provide very high starting torque, hence providing instant power once loaded. A pilot plant is moored in the Strait of Messina, Sicilian shore in Italy with the average tidal current of 2 m/s. A floating platform is above the surface of the water and is readily accessible for maintenance and repair of the turbine. The system can produce power of 20 kW. Fig. 14. Kobold turbine (left side) and floating platform (right side) Blue Energy tidal turbine consists of four fixed hydrofoil blades connected to a rotor that drives an integrated gearbox and electrical generator as shown in Fig. 15. The turbine is mounted in a durable concrete marine caisson which anchors the unit to the seabed. The platform is above the surface of the water and is readily accessible for maintenance and repair. The rotation of the turbine is unidirectional on both the ebb and the flow of the tide. A unit turbine is expected to be about 200 kW output power. For large scale power production, multiple turbines are linked in series to create a tidal fence across an ocean passage or inlet. Fig. 15. Blue Energy tidal turbine 4.3 Oscillating devices – Stingray tidal turbine Stingray uses the flow of the tidal stream over a hydroplane to create an oscillating motion that operates hydraulic cylinders to drive a motor that, in turn, drives an electrical generator. This device is a seabed mounted machine, to be situated typically in any water depth up to 100m. During 2003, a 150kW Stingray was tested off the Scottish coast, Marine Tidal Current Electric Power Generation: State of Art and Current Status 221Renewable Energy220 Fig. 12. Gorlov turbine (Source: http://www.oce.uri.edu/oce311/) Neptune Renewable Energy has developed Proteus (Neptune, 2009). The Neptune Proteus is a vertical axis turbine mounted within a symmetrical Venturi diffuser duct and beneath a very simple steel deck and buoyancy chambers. The Neptune Proteus is designed for estuarine sites that have high tidal currents, yet have the advantages of lower access, cabling and maintenance costs than in offshore environments. The vertical shaft connects to the 1:200 gearbox and generator in the deck housing. The device is moored in the free stream, minimising environmental impact and operates equally efficiently for both flood and ebb currents. The rotor is maintained at optimal power outputs by sets of computer controlled shutters within the duct and by the variable electrical load. The overall efficiency of the system suggests to be greater than 45%. Fig. 13. The Neptune Proteus The Enermar Project developed Kobold turbine (Ponte, 2009) that can provide very high starting torque, hence providing instant power once loaded. A pilot plant is moored in the Strait of Messina, Sicilian shore in Italy with the average tidal current of 2 m/s. A floating platform is above the surface of the water and is readily accessible for maintenance and repair of the turbine. The system can produce power of 20 kW. Fig. 14. Kobold turbine (left side) and floating platform (right side) Blue Energy tidal turbine consists of four fixed hydrofoil blades connected to a rotor that drives an integrated gearbox and electrical generator as shown in Fig. 15. The turbine is mounted in a durable concrete marine caisson which anchors the unit to the seabed. The platform is above the surface of the water and is readily accessible for maintenance and repair. The rotation of the turbine is unidirectional on both the ebb and the flow of the tide. A unit turbine is expected to be about 200 kW output power. For large scale power production, multiple turbines are linked in series to create a tidal fence across an ocean passage or inlet. Fig. 15. Blue Energy tidal turbine 4.3 Oscillating devices – Stingray tidal turbine Stingray uses the flow of the tidal stream over a hydroplane to create an oscillating motion that operates hydraulic cylinders to drive a motor that, in turn, drives an electrical generator. This device is a seabed mounted machine, to be situated typically in any water depth up to 100m. During 2003, a 150kW Stingray was tested off the Scottish coast, Marine Tidal Current Electric Power Generation: State of Art and Current Status 221Renewable Energy222 including a flexible control system to allow the performance of the generator to be accurately controlled and recorded over a longer period (IHC, 2009). Fig. 16. Stingray tidal energy conversion device using oscillating hydroplane (Source: http://www.engb.com/services_09a.php) 4.4 Shrouded Tidal Turbines A Patented tidal stream turbine invented by Aaron Davidson and Craig Hill of Tidal Energy Pty Ltd, Australian company, uses a venturi shaped shroud to increase the flow rate through the turbine, hence increasing the turbine efficiency as much as 3.84 times compared to the same turbine without the shroud. The Tidal Energy installed and tested such shrouded tidal turbines on the Gold Coast, Queensland in 2002. Tidal Energy has installed another shrouded turbine for a remote Australian community in northern Australia, providing 3.5 MW of power to the community. Another larger 5 meter diameter turbine, capable of 800 kW in 4 m/s of tidal flow, is planned for deployment as a tidal powered desalination showcase near Brisbane Australia in October 2008 (Hirch, 2009). Another device, the Hydro Venturi, is to be tested in San Francisco Bay (Hammons, 2008). Fig. 17. Tidal turbine with venturi shroud (Source: http://www.reuk.co.uk/Lunar-Energy-Tidal-Power.htm) In late April 2008, Ocean Renewable Power Company, LLC (ORPC) (EMEC, 2009) successfully completed the testing of its turbine-generator unit (TGU) prototype at ORPC’s Cobscook Bay and Western Passage tidal sites near Eastport, Maine (Viscarolasaga, 2008). The TGU makes use of cross-flow (ADCF) turbines to drive a permanent magnet generator located between the turbines and mounted on the same shaft. This technology can be used for generating power from river, tidal and deep water ocean currents. 4.5 Current Activities for Commercial Developments of Tidal Technologies RWE's npower announced that it is in partnership with Marine Current Turbines to build a tidal farm of SeaGen turbines off the coast of Anglesey in Wales (MCT, 2008). In November 2007, British company Lunar Energy announced that, in conjunction with E.ON, they would be building the world's first tidal energy farm off the coast of Pembrokshire in Wales. It will be the world's first deep-sea tidal-energy farm and will provide electricity for 5,000 homes. Eight underwater turbines, each 25 metres long and 15 metres high, are to be installed on the sea bottom off St David's peninsula. Construction is due to start in the summer of 2008 and the proposed tidal energy turbines, described as "a wind farm under the sea", should be operational by 2010. British Columbia Tidal Energy Corp. plans to deploy at least three 1.2 MW turbines in the Campbell River or in the surrounding coastline of British Columbia by 2009 (Alternative Energy Press, 2007). An organisation, named Alderney Renewable Energy Ltd, is planning to use tidal turbines to extract power from the notoriously strong tidal races around Alderney in the Channel Islands. It is estimated that up to 3GW could be extracted. This would not only supply the island's needs but also leave a considerable surplus for export (Alderney Renewable Energy, 2009). Nova Scotia Power has selected OpenHydro's turbine for a tidal energy demonstration project in the Bay of Fundy, Nova Scotia, Canada and Alderney Renewable Energy Ltd for the supply of tidal turbines in the Channel Islands. Open Hydro. 5. Conclusion Tidal energy is a promising renewable energy source available to the world. Since the past one decade, numerous research and development efforts have been carried out hoping that tidal energy can become a realistic renewable energy source one day. At present, tidal current technologies are still in the developing stage. A large variety of tidal designs have been developed and experimented. Several prototypes have been claimed to be promising. Some companies are planning for large scale manufacturing and deployment of tidal turbine. Any commercial tidal plants are still in the planning stage. 6. References Alderney Renewable Energy. (2009). Technology Overview. Source: http:// www.are.gb.com/technology-overview.php. Access: 30th March 2009 Marine Tidal Current Electric Power Generation: State of Art and Current Status 223Renewable Energy222 including a flexible control system to allow the performance of the generator to be accurately controlled and recorded over a longer period (IHC, 2009). Fig. 16. Stingray tidal energy conversion device using oscillating hydroplane (Source: http://www.engb.com/services_09a.php) 4.4 Shrouded Tidal Turbines A Patented tidal stream turbine invented by Aaron Davidson and Craig Hill of Tidal Energy Pty Ltd, Australian company, uses a venturi shaped shroud to increase the flow rate through the turbine, hence increasing the turbine efficiency as much as 3.84 times compared to the same turbine without the shroud. The Tidal Energy installed and tested such shrouded tidal turbines on the Gold Coast, Queensland in 2002. Tidal Energy has installed another shrouded turbine for a remote Australian community in northern Australia, providing 3.5 MW of power to the community. Another larger 5 meter diameter turbine, capable of 800 kW in 4 m/s of tidal flow, is planned for deployment as a tidal powered desalination showcase near Brisbane Australia in October 2008 (Hirch, 2009). Another device, the Hydro Venturi, is to be tested in San Francisco Bay (Hammons, 2008). Fig. 17. Tidal turbine with venturi shroud (Source: http://www.reuk.co.uk/Lunar-Energy-Tidal-Power.htm) In late April 2008, Ocean Renewable Power Company, LLC (ORPC) (EMEC, 2009) successfully completed the testing of its turbine-generator unit (TGU) prototype at ORPC’s Cobscook Bay and Western Passage tidal sites near Eastport, Maine (Viscarolasaga, 2008). The TGU makes use of cross-flow (ADCF) turbines to drive a permanent magnet generator located between the turbines and mounted on the same shaft. This technology can be used for generating power from river, tidal and deep water ocean currents. 4.5 Current Activities for Commercial Developments of Tidal Technologies RWE's npower announced that it is in partnership with Marine Current Turbines to build a tidal farm of SeaGen turbines off the coast of Anglesey in Wales (MCT, 2008). In November 2007, British company Lunar Energy announced that, in conjunction with E.ON, they would be building the world's first tidal energy farm off the coast of Pembrokshire in Wales. It will be the world's first deep-sea tidal-energy farm and will provide electricity for 5,000 homes. Eight underwater turbines, each 25 metres long and 15 metres high, are to be installed on the sea bottom off St David's peninsula. Construction is due to start in the summer of 2008 and the proposed tidal energy turbines, described as "a wind farm under the sea", should be operational by 2010. British Columbia Tidal Energy Corp. plans to deploy at least three 1.2 MW turbines in the Campbell River or in the surrounding coastline of British Columbia by 2009 (Alternative Energy Press, 2007). An organisation, named Alderney Renewable Energy Ltd, is planning to use tidal turbines to extract power from the notoriously strong tidal races around Alderney in the Channel Islands. It is estimated that up to 3GW could be extracted. This would not only supply the island's needs but also leave a considerable surplus for export (Alderney Renewable Energy, 2009). Nova Scotia Power has selected OpenHydro's turbine for a tidal energy demonstration project in the Bay of Fundy, Nova Scotia, Canada and Alderney Renewable Energy Ltd for the supply of tidal turbines in the Channel Islands. Open Hydro. 5. Conclusion Tidal energy is a promising renewable energy source available to the world. Since the past one decade, numerous research and development efforts have been carried out hoping that tidal energy can become a realistic renewable energy source one day. At present, tidal current technologies are still in the developing stage. A large variety of tidal designs have been developed and experimented. Several prototypes have been claimed to be promising. Some companies are planning for large scale manufacturing and deployment of tidal turbine. Any commercial tidal plants are still in the planning stage. 6. References Alderney Renewable Energy. (2009). Technology Overview. Source: http:// www.are.gb.com/technology-overview.php. Access: 30th March 2009 Marine Tidal Current Electric Power Generation: State of Art and Current Status 223Renewable Energy224 Alternative Energy Press. (2007). Tidal power coming to west coast of Canada. Source: http://www.alternative-energy-news.info/press/tidal-power-west-coast-canada/. (8th November 2007) Bahaj, A. S.; & Myers, L. (2004). Analytical estimates of the energy yield potential from the Alderney Race (Channel Islands) using current energy converters. Renewable Energy, 29, 12, pp. 1931-1945 (2004), ISSN: 0960-1481 Bellve, Anthony. (2005). NZ: Chance to turn the tide of power supply. New Zealand Herald, (9th May 2005). Available: http://www.nzherald.co.nz/nz/news/ article.cfm?c_id=1&objectid=10124626 Cascio, Jamais. (2005). Gorlov's Helical Turbine. WorldChanging, Available: http://www.worldchanging.com/archives/002383.html, (21st March 2005) CBC News. (2007). The resurgence of tidal power. Source: http://www.cbc.ca/news/ background/energy/tidal-power.html. 26th February 2007 Charlie, R. H. (2007). Forty Candles for the Rance River TPP tides provide renewable and sustainable power generation. Renewable and Sustainable Energy Reviews, 11, 9, (December 2007) pp. 2032-2057, ISSN: 1364-0321 Deng, Zhiqun; Carlson, Thomas J.; Ploskey, Gene R.; Richmond, Marshall C.; & Dauble, Dennis D. (2007). Evaluation of blade-strike models for estimating the biological performance of Kaplan turbines. Ecological Modelling, 208, 2-4, (November 2007) pp. 165-176, ISSN: 0304-3800 Eling (2009). Eling Tide Mill. Source : www.elingtidemill.org.uk, Date of Access : 25th March 2009 EMEC. (2008). Tidal Devices. Source : European Marine Energy Centre (EMEC) Ltd (http://www.emec.org.uk/tidal_devices.asp), Date of Access : 25th March 2009 European Commission. (2008). United Kingdom renewable energy fact sheet. Source : http://www.energy.eu/renewables/factsheets/2008_res_sheet_united_kingdom_en.pdf. January 2008 Fraenkel, Peter. (2004). Windmills below the sea: A commercial reality soon?. Refocus, 5, 2 (March-April 2004), pp. 46-48, ISSN : 1471-0846 Galbraith, Kate. (2008). Power From the Restless Sea Stirs the Imagination. The New York Times, (23 September 2008), Available : http://www.nytimes.com/2008/09/23/ business/23tidal.html?_r=1&em Gevorkian, Peter. (1999). Ocean Energy Technology, In: Sustainable energy systems engineering: the complete green building design resource, Peter Gevorkian, (1st Ed.), pp. 255-272, McGraw-Hill Professional, ISBN 0071473599, 9780071473590, New York, USA Hagerman, G. (2006). Tidal stream energy in the Bay of Fundy. Proceedings of Energy Research and Development Forum 2006 Antigonish, Nova Scotia, Canada, 2006. Hammons, T. J. (2008). Energy potential of the oceans in Europe and North America: tidal, wave, currents, OTEC and offshore wind. International Journal of Power & Energy Systems, 28, pp. 416-428 (2008), ISSN: 1078-3466 Hardisty, Jack. (2007). Power intermittency, redundancy and tidal phasing around the United Kingdom. Geographical Journal, 174, 1, pp. 76-84 (December 2007), ISSN: 0373-6245 Hirch, Matthew. (2009). Low tide - Without state or federal funding, S.F.'s tidal power project might be dead in the water. Sanfrancisco Bay Guardian Online. Available: http://www.sfbg.com/38/43/news_tidal.html. Access: 30th March 2009. IHC. (2009). Stingray tidal stream generator. Source: http://www.engb.com/ services_09a.php. Access: 30th March 2009 Livingstone, Tomos. (2007). Severn balancing act – Hain. WalesOnline.co.uk, (1st September 2007). Available: http://www.walesonline.co.uk/news/welsh-politics/welsh-politics-news/tm_headline=severn-balancing-act-hain&method=full&objectid=19718602&siteid=50082-name_page.html McDermott, Matthew. (2009). 1.2MW DeltaStream tidal power turbines to be tested in Wales. Science & Technology, (5th January 2009). Available: http:// www.treehugger.com/files/2009/01/delta-stream-tidal-power-turbine-wales.php MCT. (2008). World's first commercial-scale tidal power system feeds electricity to the National Grid. Source: http://www.marineturbines.com/3/news/article/10/ world_s_first_commercial_scale_tidal_power_system_feeds_electricity_to_the_national_grid__/. 17th July 2008 Neptune. (2009). Tidal Technology – The Neptune Proteus Mark III. Source: http://www.neptunerenewableenergy.com/tidal_technology.php. Access: 25th March 2009 Pelc, Robin.; & Fujita, Rod M. (2002). Renewable energy from the ocean. Marine Policy, 26, 6, (November 2002) pp. 471-479, ISSN: 0308-597x Peltier, Robert. (2003). Fish-friendly hydro turbines move center stage. Energy Industry, April 2003. Available: http://findarticles.com/p/articles/mi_qa5392/is_200304/ ai_n21329129/ Penman, Danny. (2003). First power station to harness Moon opens. Source : http:// www.newscientist.com/article/dn4188, November 2003 Ponte (2009). Ponte di Archimede International Company Source: http:// www.pontediarchimede.coi/ (April 2009). Retiere, C. (1994). Tidal power and aquatic environment of La Rance. Biological Journal of the Linnean Society, 51, 1-2, (January 1993) pp. 25-36, ISSN: 0024-4066 Schmitt, Catherine. (2009). Tidal power proposals proliferate along Maine's coast. The Working Waterfront, 2nd April 2009. Available: http://www.workingwaterfront.com /articles/Tidal-power-proposals-proliferate-along-Maines-coast/11432/ The Renewable Energy Centre. (2008). The Renewable Energy Centre supports use of Tidal power to supply the National Grid. Source: http://www.prlog.org/10076285-the-renewable-energy-centre-supports-use-of-tidal-power-to-supply-the-national-grid.html. 30th May 2008 Tom, G. (2007), Oceanography : An Invitation to Marine Science, Thomson Brooks/Cole, Orange Coast College, University of Southern California Urbina, Ian. (2004). In Search of New Power Source, City Looks Underwater. The New York Times, pp B3 (10th July 2004) Vega, Cecilia M. (2007). PG&E backs new study of bay's tidal power. Source : http://deanzaemtp.googlepages.com/PGEbacksnewstudyofbaystidalpower.pdf. (20th June 2007) Verdant Power. (2009). The RITE project. Source : http://www.verdantpower.com/what-initiative, Date of Access : 25th March 2009 Marine Tidal Current Electric Power Generation: State of Art and Current Status 225Renewable Energy224 Alternative Energy Press. (2007). Tidal power coming to west coast of Canada. Source: http://www.alternative-energy-news.info/press/tidal-power-west-coast-canada/. (8th November 2007) Bahaj, A. S.; & Myers, L. (2004). Analytical estimates of the energy yield potential from the Alderney Race (Channel Islands) using current energy converters. Renewable Energy, 29, 12, pp. 1931-1945 (2004), ISSN: 0960-1481 Bellve, Anthony. (2005). NZ: Chance to turn the tide of power supply. New Zealand Herald, (9th May 2005). Available: http://www.nzherald.co.nz/nz/news/ article.cfm?c_id=1&objectid=10124626 Cascio, Jamais. (2005). Gorlov's Helical Turbine. WorldChanging, Available: http://www.worldchanging.com/archives/002383.html, (21st March 2005) CBC News. (2007). The resurgence of tidal power. Source: http://www.cbc.ca/news/ background/energy/tidal-power.html. 26th February 2007 Charlie, R. H. (2007). Forty Candles for the Rance River TPP tides provide renewable and sustainable power generation. Renewable and Sustainable Energy Reviews, 11, 9, (December 2007) pp. 2032-2057, ISSN: 1364-0321 Deng, Zhiqun; Carlson, Thomas J.; Ploskey, Gene R.; Richmond, Marshall C.; & Dauble, Dennis D. (2007). Evaluation of blade-strike models for estimating the biological performance of Kaplan turbines. Ecological Modelling, 208, 2-4, (November 2007) pp. 165-176, ISSN: 0304-3800 Eling (2009). Eling Tide Mill. Source : www.elingtidemill.org.uk, Date of Access : 25th March 2009 EMEC. (2008). Tidal Devices. Source : European Marine Energy Centre (EMEC) Ltd (http://www.emec.org.uk/tidal_devices.asp), Date of Access : 25th March 2009 European Commission. (2008). United Kingdom renewable energy fact sheet. Source : http://www.energy.eu/renewables/factsheets/2008_res_sheet_united_kingdom_en.pdf. January 2008 Fraenkel, Peter. (2004). Windmills below the sea: A commercial reality soon?. Refocus, 5, 2 (March-April 2004), pp. 46-48, ISSN : 1471-0846 Galbraith, Kate. (2008). Power From the Restless Sea Stirs the Imagination. The New York Times, (23 September 2008), Available : http://www.nytimes.com/2008/09/23/ business/23tidal.html?_r=1&em Gevorkian, Peter. (1999). Ocean Energy Technology, In: Sustainable energy systems engineering: the complete green building design resource, Peter Gevorkian, (1st Ed.), pp. 255-272, McGraw-Hill Professional, ISBN 0071473599, 9780071473590, New York, USA Hagerman, G. (2006). Tidal stream energy in the Bay of Fundy. Proceedings of Energy Research and Development Forum 2006 Antigonish, Nova Scotia, Canada, 2006. Hammons, T. J. (2008). Energy potential of the oceans in Europe and North America: tidal, wave, currents, OTEC and offshore wind. International Journal of Power & Energy Systems, 28, pp. 416-428 (2008), ISSN: 1078-3466 Hardisty, Jack. (2007). Power intermittency, redundancy and tidal phasing around the United Kingdom. Geographical Journal, 174, 1, pp. 76-84 (December 2007), ISSN: 0373-6245 Hirch, Matthew. (2009). Low tide - Without state or federal funding, S.F.'s tidal power project might be dead in the water. Sanfrancisco Bay Guardian Online. Available: http://www.sfbg.com/38/43/news_tidal.html. Access: 30th March 2009. IHC. (2009). Stingray tidal stream generator. Source: http://www.engb.com/ services_09a.php. Access: 30th March 2009 Livingstone, Tomos. (2007). Severn balancing act – Hain. WalesOnline.co.uk, (1st September 2007). Available: http://www.walesonline.co.uk/news/welsh-politics/welsh-politics-news/tm_headline=severn-balancing-act-hain&method=full&objectid=19718602&siteid=50082-name_page.html McDermott, Matthew. (2009). 1.2MW DeltaStream tidal power turbines to be tested in Wales. Science & Technology, (5th January 2009). Available: http:// www.treehugger.com/files/2009/01/delta-stream-tidal-power-turbine-wales.php MCT. (2008). World's first commercial-scale tidal power system feeds electricity to the National Grid. Source: http://www.marineturbines.com/3/news/article/10/ world_s_first_commercial_scale_tidal_power_system_feeds_electricity_to_the_national_grid__/. 17th July 2008 Neptune. (2009). Tidal Technology – The Neptune Proteus Mark III. Source: http://www.neptunerenewableenergy.com/tidal_technology.php. Access: 25th March 2009 Pelc, Robin.; & Fujita, Rod M. (2002). Renewable energy from the ocean. Marine Policy, 26, 6, (November 2002) pp. 471-479, ISSN: 0308-597x Peltier, Robert. (2003). Fish-friendly hydro turbines move center stage. Energy Industry, April 2003. Available: http://findarticles.com/p/articles/mi_qa5392/is_200304/ ai_n21329129/ Penman, Danny. (2003). First power station to harness Moon opens. Source : http:// www.newscientist.com/article/dn4188, November 2003 Ponte (2009). Ponte di Archimede International Company Source: http:// www.pontediarchimede.coi/ (April 2009). Retiere, C. (1994). Tidal power and aquatic environment of La Rance. Biological Journal of the Linnean Society, 51, 1-2, (January 1993) pp. 25-36, ISSN: 0024-4066 Schmitt, Catherine. (2009). Tidal power proposals proliferate along Maine's coast. The Working Waterfront, 2nd April 2009. Available: http://www.workingwaterfront.com /articles/Tidal-power-proposals-proliferate-along-Maines-coast/11432/ The Renewable Energy Centre. (2008). The Renewable Energy Centre supports use of Tidal power to supply the National Grid. Source: http://www.prlog.org/10076285-the-renewable-energy-centre-supports-use-of-tidal-power-to-supply-the-national-grid.html. 30th May 2008 Tom, G. (2007), Oceanography : An Invitation to Marine Science, Thomson Brooks/Cole, Orange Coast College, University of Southern California Urbina, Ian. (2004). In Search of New Power Source, City Looks Underwater. The New York Times, pp B3 (10th July 2004) Vega, Cecilia M. (2007). PG&E backs new study of bay's tidal power. Source : http://deanzaemtp.googlepages.com/PGEbacksnewstudyofbaystidalpower.pdf. (20th June 2007) Verdant Power. (2009). The RITE project. Source : http://www.verdantpower.com/what-initiative, Date of Access : 25th March 2009 Marine Tidal Current Electric Power Generation: State of Art and Current Status 225Renewable Energy226 Viscarolasaga, Efrain. (2008). Tide is slowly rising in interest in ocean power. Mass High Tech: The Journal of New England Technology, (1st August 2008) Available : http://www.masshightech.com/stories/2008/07/28/weekly9-Tide-is-slowly-rising-in-interest-in-ocean-power.html/ VLH (2009). Very low head turbine. Source : http://www.vlh-turbine.com/, Date of Access : 25th March 2009 Modelling and Simulation of an Induction Drive with Application to a Small Wind Turbine Generator 227Modelling and Simulation of an Induction Drive with Application to a Small Wind Turbine GeneratorLevente TAMAS and Zoltan SZEKELYX Modeling and Simulation of an Induction Drive with Application to a Small Wind Turbine Generator Levente TAMAS1 and Zoltan SZEKELY2 1Technical University of Cluj-Napoca, Romania 2Purdue University Calumet, USA 1. Introduction In recent years the rapid development of digital signal processors (DSP) based systems and the decreasing cost of power electronics allowed the complex models and control algorithms of AC machines to become popular for a large range of applications including the ones for renewable energies. Among renewable energies the solution of utilizing induction generators for wind energy conversion systems is now in a growing trend. Currently, a wide spread control concept is that of a variable speed rotor with pitch regulation, and this concept is combined with both direct drive and geared drive trains the latter dominating the wind market (Tamas & Szekely, 2008). A valid choice for operation at variable speed may be the use of induction machine. An induction motor can operate as a generator in super synchronous speed raised by an overhauling type of load, or by lowering the inverter frequency below the machine speed, when there is a converter-fed machine drive. Continuous regenerative operation of a drive is possible if the load machine is a source of power, such as in a wind generation system (Tamas & Szekely, 2008). In this paper the model of the induction machine is written in terms of fluxes as state variables. This model is then used to design an Extended Kalman Filter (EKF) based estimator that can provide the correct feedback signals for the vector control (VC) of a wind energy conversion system. Also three other observers are shown with good simulation results, one based on stator voltages and stator currents (Us-Is), one based on stator voltages and rotor speed (Us-ω) and one based on stator currents and rotor speed (Is – ω). A neural network based speed estimator as a forth type of estimator is developed which can be used for sensorless vector control of the induction drive or as a part of a sensor fusion based feedback signal estimator combined with one of the above mentioned estimators. Finally, a real-time simulation setup is presented using dSpace CLP1104 embedded system (Tamas & Szekely, 2008) together with the possible further research steps in this project. 13Renewable Energy226 Viscarolasaga, Efrain. (2008). Tide is slowly rising in interest in ocean power. Mass High Tech: The Journal of New England Technology, (1st August 2008) Available : http://www.masshightech.com/stories/2008/07/28/weekly9-Tide-is-slowly-rising-in-interest-in-ocean-power.html/ VLH (2009). Very low head turbine. Source : http://www.vlh-turbine.com/, Date of Access : 25th March 2009 Modelling and Simulation of an Induction Drive with Application to a Small Wind Turbine Generator 227Modelling and Simulation of an Induction Drive with Application to a Small Wind Turbine GeneratorLevente TAMAS and Zoltan SZEKELYX Modeling and Simulation of an Induction Drive with Application to a Small Wind Turbine Generator Levente TAMAS1 and Zoltan SZEKELY2 1Technical University of Cluj-Napoca, Romania 2Purdue University Calumet, USA 1. Introduction In recent years the rapid development of digital signal processors (DSP) based systems and the decreasing cost of power electronics allowed the complex models and control algorithms of AC machines to become popular for a large range of applications including the ones for renewable energies. Among renewable energies the solution of utilizing induction generators for wind energy conversion systems is now in a growing trend. Currently, a wide spread control concept is that of a variable speed rotor with pitch regulation, and this concept is combined with both direct drive and geared drive trains the latter dominating the wind market (Tamas & Szekely, 2008). A valid choice for operation at variable speed may be the use of induction machine. An induction motor can operate as a generator in super synchronous speed raised by an overhauling type of load, or by lowering the inverter frequency below the machine speed, when there is a converter-fed machine drive. Continuous regenerative operation of a drive is possible if the load machine is a source of power, such as in a wind generation system (Tamas & Szekely, 2008). In this paper the model of the induction machine is written in terms of fluxes as state variables. This model is then used to design an Extended Kalman Filter (EKF) based estimator that can provide the correct feedback signals for the vector control (VC) of a wind energy conversion system. Also three other observers are shown with good simulation results, one based on stator voltages and stator currents (Us-Is), one based on stator voltages and rotor speed (Us-ω) and one based on stator currents and rotor speed (Is – ω). A neural network based speed estimator as a forth type of estimator is developed which can be used for sensorless vector control of the induction drive or as a part of a sensor fusion based feedback signal estimator combined with one of the above mentioned estimators. Finally, a real-time simulation setup is presented using dSpace CLP1104 embedded system (Tamas & Szekely, 2008) together with the possible further research steps in this project. 13Renewable Energy228 2. Wind Turbine System Model Fig. 1 shows a basic wind turbine system that converts the wind energy into mechanical energy which at its turn is used by an electric machine to generate power. The basic configuration consists of the wind turbine blades, the drive train, the induction generator and the AC-AC converter (not shown in Fig. 1 to connect the system to the grid). Fig. 1. Basic wind turbine system model 2.1 Wind turbine model The models of wind turbines take into account several characteristics such as the size, blade radius, nominal power, shafts stiffness, losses, gear box ratio, etc. The mechanical power produced by the wind turbine is: 3ν2πRpcairρ21MP  (1) where  is the air density, 2R is the turbine swept area, v is the speed of the wind, and cp(λ) is the power efficiency coefficient of the wind turbine which depends on the tip-speed ratio, λ. This is defined as: νRMωλ  (2) where, dθMωM dt is the angular speed of the turbine blades, and R is the length of the turbine blade. There is an optimal value of the tip speed ratio, λopt, which allows a maximum capture of power from the wind. This value is found from the typical characteristic of the wind turbine power coefficient cp = f(λ) provided by the manufacturers. When the wind turbine is operating with λopt, the nonlinear power expression from (1) may be recalculated as: 3MωoptcoptMP  (3) and the corresponding mechanical torque produced by the blades is: 2MωoptcoptMT  (4) where, Copt is a constant which depends on the turbine characteristics and air density (Kelemen & Imecs, 1991). 2.2 Drive train model The drive train dynamics consists of the dynamics of the wind turbine rotor, low-speed shaft, gear box, high-speed shaft, and the induction generator Fig. 2 shows the basic diagram of the drive train that was used for the derivation of the mathematical model (Kelemen & Imecs, 1991). Thus, for the wind turbine rotor the following torques equation can be written: (t)MTGT2dtMθ2dMJ  (5) where, T K ΔθG 1 K1 is the reaction torque exerted by the flexible low-speed shaft on the rotor of the wind turbine. Similarly, for the rotor of the induction generator the following torques equation is valid: 0eTgT2dtrθ2dgJ  (6) where T K Δθg 2 K2 is the reaction torque exerted by the high-speed shaft on the induction generator rotor, and eT is the electromagnetic torque of the induction machine in generator operation mode. Fig. 2. Drive train model The reference values for the above mechanical parameters where reported in (Scutaru & Apostoaia, 2004). Modelling and Simulation of an Induction Drive with Application to a Small Wind Turbine Generator 229Renewable Energy228 2. Wind Turbine System Model Fig. 1 shows a basic wind turbine system that converts the wind energy into mechanical energy which at its turn is used by an electric machine to generate power. The basic configuration consists of the wind turbine blade

How to achieve 100% renewable energy

Abstract

Extracted data

How to achieve 100% renewable energy

Abstract

Extracted data

Related items

Related items