Emilia Motoasca Avinash Kumar Agarwal Hilde Breesch Editors Energy Sustainability in Built and Urban Environments Energy, Environment, and Sustainability

Energy, Environment, and Sustainability Series Editors: Avinash Kumar Agarwal · Ashok Pandey

Series editors Avinash Kumar Agarwal, Department of Mechanical Engineering, Indian Institute of Technology Kanpur, Kanpur, Uttar Pradesh, India Ashok Pandey, Distinguished Scientist, CSIR-Indian Institute of Toxicology Research, Lucknow, Uttar Pradesh, India This books series publishes cutting edge monographs and professional books focused on all aspects of energy and environmental sustainability, especially as it relates to energy concerns. The Series is published in partnership with the International Society for Energy, Environment, and Sustainability. The books in these series are editor or authored by top researchers and professional across the globe. The series aims at publishing state-of-the-art research and development in areas including, but not limited to: • Renewable Energy • Alternative Fuels • Engines and Locomotives • Combustion and Propulsion • Fossil Fuels • Carbon Capture • Control and Automation for Energy • Environmental Pollution • Waste Management • Transportation Sustainability

More information about this series at http://www.springer.com/series/15901 Emilia Motoasca • Avinash Kumar Agarwal Hilde Breesch Editors

Energy Sustainability in Built and Urban Environments

123 Editors Emilia Motoasca Hilde Breesch Department of Electrical Engineering Department of Civil Engineering KU Leuven (Catholic University Leuven) KU Leuven (Catholic University Leuven) Ghent, Belgium Ghent, Belgium

Avinash Kumar Agarwal Department of Mechanical Engineering Indian Institute of Technology Kanpur Kanpur, Uttar Pradesh, India

ISSN 2522-8366 ISSN 2522-8374 (electronic) Energy, Environment, and Sustainability ISBN 978-981-13-3283-8 ISBN 978-981-13-3284-5 (eBook) https://doi.org/10.1007/978-981-13-3284-5

Library of Congress Control Number: 2018961716

© Springer Nature Singapore Pte Ltd. 2019 This work is subject to copyright. All rights are reserved by the Publisher, whether the whole or part of the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations, recitation, broadcasting, reproduction on microfilms or in any other physical way, and transmission or information storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar methodology now known or hereafter developed. The use of general descriptive names, registered names, trademarks, service marks, etc. in this publication does not imply, even in the absence of a specific statement, that such names are exempt from the relevant protective laws and regulations and therefore free for general use. The publisher, the authors and the editors are safe to assume that the advice and information in this book are believed to be true and accurate at the date of publication. Neither the publisher nor the authors or the editors give a warranty, express or implied, with respect to the material contained herein or for any errors or omissions that may have been made. The publisher remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

This Springer imprint is published by the registered company Springer Nature Singapore Pte Ltd. The registered company address is: 152 Beach Road, #21-01/04 Gateway East, Singapore 189721, Singapore Preface

Energy demand has been rising remarkably due to increasing population and urbanization. Global economy and society are significantly dependent on the energy availability because it touches every facet of human life and its activities. Transportation and power generation are two major examples. Without the transportation by millions of personalized and mass transport vehicles and availability of 24 Â 7 power, human civilization would not have reached contemporary living standards. The International Society for Energy, Environment and Sustainability (ISEES) was founded at Indian Institute of Technology Kanpur (IIT Kanpur), India, in January 2014 with the aim of spreading knowledge/awareness and catalysing research activities in the fields of energy, environment, sustainability and combustion. The society’s goal is to contribute to the development of clean, affordable and secure energy resources and a sustainable environment for the society and to spread knowledge in the above-mentioned areas and create awareness about the environmental challenges, which the world is facing today. The unique way adopted by the society was to break the conventional silos of specializations (engineering, science, environment, agriculture, biotechnology, materials, fuels, etc.) to tackle the problems related to energy, environment and sustainability in a holistic manner. This is quite evident by the participation of experts from all fields to resolve these issues. ISEES is involved in various activities such as conducting workshops, seminars and conferences in the domains of its interest. The society also recognizes the outstanding works done by the young scientists and engineers for their contributions in these fields by conferring them awards under various categories. The second international conference on “Sustainable Energy and Environmental Challenges” (SEEC-2018) was organized under the auspices of ISEES from 31 December 2017 to 3 January 2018 at J N Tata Auditorium, Indian Institute of Science Bangalore. This conference provided a platform for discussions between eminent scientists and engineers from various countries including India, USA, South Korea, Norway, Finland, Malaysia, Austria, Saudi Arabia and Australia. In this conference, eminent speakers from all over the world presented their views

v vi Preface related to different aspects of energy, combustion, emissions and alternative energy resources for sustainable development and a cleaner environment. The conference presented five high-voltage plenary talks from globally renowned experts on topical themes, namely “Is It Really the End of Combustion Engines and Petroleum?” by Prof. Gautam Kalghatgi, Saudi Aramco; “Energy Sustainability in India: Challenges and Opportunities” by Prof. Baldev Raj, NIAS Bangalore; “Methanol Economy: An Option for Sustainable Energy and Environmental Challenges” by Dr. Vijay Kumar Saraswat, Hon. Member (S&T), NITI Aayog, Government of India; “Supercritical Carbon Dioxide Brayton Cycle for Power Generation” by Prof. Pradip Dutta, IISc Bangalore; and “Role of Nuclear Fusion for Environmental Sustainability of Energy in Future” by Prof. J. S. Rao, Altair Engineering. The conference included 27 technical sessions on topics related to energy and environmental sustainability including 5 plenary talks, 40 keynote talks and 18 invited talks from prominent scientists, in addition to 142 contributed talks, and 74 poster presentations by students and researchers. The technical sessions in the conference included Advances in IC Engines: SI Engines, Solar Energy: Storage, Fundamentals of Combustion, Environmental Protection and Sustainability, Environmental Biotechnology, Coal and Biomass Combustion/Gasification, Air Pollution and Control, Biomass to Fuels/Chemicals: Clean Fuels, Advances in IC Engines: CI Engines, Solar Energy: Performance, Biomass to Fuels/Chemicals: Production, Advances in IC Engines: Fuels, Energy Sustainability, Environmental Biotechnology, Atomization and Sprays, Combustion/Gas Turbines/Fluid Flow/Sprays, Biomass to Fuels/Chemicals, Advances in IC Engines: New Concepts, Energy Sustainability, Waste to Wealth, Conventional and Alternate Fuels, Solar Energy, Wastewater Remediation and Air Pollution. One of the highlights of the conference was the rapid-fire poster sessions in (i) Energy Engineering, (ii) Environment and Sustainability and (iii) Biotechnology, where more than 75 students participated with great enthusiasm and won many prizes in a fiercely competitive environment. More than 200 participants and speakers attended this four-day conference, which also hosted Dr. Vijay Kumar Saraswat, Hon. Member (S&T), NITI Aayog, Government of India, as the chief guest for the book release ceremony, where 16 ISEES books published by Springer, under a special dedicated series Energy, Environment, and Sustainability were released. This is the first time that such significant and high-quality outcome has been achieved by any society in India. The conference concluded with a panel discussion on “Challenges, Opportunities and Directions for Future Transportation Systems”, where the pan- ellists were Prof. Gautam Kalghatgi, Saudi Aramco; Dr. Ravi Prashanth, Caterpillar Inc.; Dr. Shankar Venugopal, Mahindra and Mahindra; Dr. Bharat Bhargava, DG, ONGC Energy Centre; and Dr. Umamaheshwar, GE Transportation, Bangalore. The panel discussion was moderated by Prof. Ashok Pandey, Chairman, ISEES. This conference laid out the road map for technology development, opportunities and challenges in energy, environment and sustainability domains. All these topics are very relevant for the country and the world in the present context. We acknowledge the support received from various funding agencies and organizations for the successful conduct of the second ISEES conference SEEC-2018, where Preface vii these books germinated. We would therefore like to acknowledge SERB, Government of India (special thanks to Dr. Rajeev Sharma, Secretary); ONGC Energy Centre (special thanks to Dr. Bharat Bhargava); TAFE (special thanks to Sh. Anadrao Patil); Caterpillar (special thanks to Dr. Ravi Prashanth); Progress Rail, TSI, India (special thanks to Dr. Deepak Sharma); Tesscorn, India (special thanks to Sh. Satyanarayana); GAIL, Volvo; and our publishing partner Springer (special thanks to Swati Meherishi). The editors would like to express their sincere gratitude to a large number of authors from all over the world for submitting their high-quality work in a timely manner and revising it appropriately at short notice. We would like to express our special thanks to Prof. Ahmed Rachid, Arch. Alexis Versele, Dr. Bart Huyck, Prof. Roger Sierens, Dr. Abhishek Dutta, Prof. Giacomo Chiesa, Dr. Ivan Korolija, Mrs. Sien Winters, Prof. Bolanle Ikotun, Mrs. Tran Thanh Vu, Prof. Pham Duc Nguyen, Mr. Bart Merema and Prof. Jos Knockaert, who reviewed various chapters of this book and provided very valuable suggestions to the authors to improve their manuscript. This book covers different aspects of energy sustainability: implementation at macro-scale (nation, city, neighbourhood) and building scale, strategies in relation to buildings, neighbourhoods, systems and energy markets and sustainable energy production, use and storage technologies. Topics include sustainable construction practices, urban planning, energy efficiency of residential, school and office buildings, how to manage the impact of future climate conditions, control strategies of microgrids and financial instruments. Wind energy, thermoelectric materials, concentrated photovoltaic, hydrogen fuel clean energy cycle and renewable energy storage are also presented here through a series of chapters.

Ghent, Belgium Emilia Motoasca Kanpur, India Avinash Kumar Agarwal Ghent, Belgium Hilde Breesch Contents

Part I Energy Sustainability Implementation 1 Sustainable Construction Practices in West African Countries .... 3 Adedayo J. Ogungbile and Ayodeji E. Oke 2 Modelling the Influence of Urban Planning on the Financial and Environmental Impact of Neighbourhoods ...... 17 Damien Trigaux, Karen Allacker and Frank De Troyer 3 Achieving Energy Efficiency in Urban Residential Buildings in Vietnam: High-tech or Low-tech? ...... 39 Quang Minh Nguyen 4 Recommendations for the Design of an Energy-Efficient and Indoor Comfortable Office Building in Vietnam ...... 67 Ngo Hoang Ngoc Dung and Nguyen Trung Kien

Part II Energy Sustainability Strategies 5 Linking Neighborhoods into Sustainable Energy Systems ...... 93 A. T. D. Perera, Silvia Coccolo, Pietro Florio, Vahid M. Nik, Dasaraden Mauree and Jean-Louis Scartezzini 6 Future Weather Data for Dynamic Building Energy Simulations: Overview of Available Data and Presentation of Newly Derived Data for Belgium ...... 111 Delphine Ramon, Karen Allacker, Nicole P. M. van Lipzig, Frank De Troyer and Hendrik Wouters 7 Evaluation of a Simpliﬁed Calculation Approach for Final Heating Energy Use in Non-residential Buildings ...... 139 Barbara Wauman, Wout Parys, Hilde Breesch and Dirk Saelens

ix x Contents

8 A Review of Fuzzy-Based Residential Grid-Connected Microgrid Energy Management Strategies for Grid Power Proﬁle Smoothing ...... 165 Diego Arcos-Aviles, Francesc Guinjoan, Julio Pascual, Luis Marroyo, Pablo Sanchis, Rodolfo Gordillo, Paúl Ayala and Martin P. Marietta 9 Analyzing Alternative Energy Mutual Fund Performance in the Spanish Market ...... 201 Carmen-Pilar Martí-Ballester

Part III Energy Sustainability Technologies 10 Wind Energy...... 217 J. Peuteman 11 Energy Sustainability Through the Use of Thermoelectric Materials in Waste Heat Recovery Systems Recent Developments and Challenges ...... 237 Emilia Motoasca 12 Optimization Strategy of Sustainable Concentrated Photovoltaic Thermal (CPVT) System for Cooling ...... 255 Muhammad Burhan, Muhammad Wakil Shahzad and Kim Choon Ng 13 Novel Method and Molten Salt Electrolytic Cell for Implementing a Hydrogen Fuel, Sustainable, Closed Clean Energy Cycle on a Large Scale ...... 277 Alvin G. Stern 14 Renewable Energy Storage and Its Application for Desalination ...... 313 Muhammad Wakil Shahzad, Muhammad Burhan and Kim Choon Ng Editors and Contributors

About the Editors

Emilia Motoasca (Ph.D.) is an assistant professor at Electrical Engineering (ESAT) TC, KU Leuven (Catholic University of Leuven), Belgium. She has previously worked as an assistant professor and postdoc in KU Leuven and Eindhoven University of Tech- nology, the Netherlands, respectively. Her research interests are in the design of electric/hydraulic drive trains, electric motors and other types of actuators; assistive devices and energy-efﬁcient design. She has authored more than 34 research papers and holds 1 patent. Avinash Kumar Agarwal (Ph.D.) is a professor in the Department of Mechanical Engineering at Indian Institute of Technology Kanpur. His areas of interest are IC engines, combustion, alternative fuels, conventional fuels, optical diagnostics, laser ignition, HCCI, emission and particulate control, and large bore engines. He has published 24 books and more than 230 international journal and conference papers. He is a fellow of SAE (2012), ASME (2013), ISEES (2015) and INAE (2015). He received several awards such as prestigious Shanti Swarup Bhatnagar Award-2016 in engineering sciences; Rajib Goyal Prize-2015; NASI-Reliance Industries Platinum Jubilee Award-2012; INAE Silver Jubilee Young Engineer Award-2012; SAE International’s

xi xii Editors and Contributors

Ralph R. Teetor Educational Award-2008; INSA Young Scientist Award-2007; UICT Young Scientist Award- 2007 and INAE Young Engineer Award-2005. Hilde Breesch (Ph.D.) is an assistant professor in the Department of Civil Engineering, KU Leuven (Catholic University Leuven), and the head of Construction Technology Cluster at Technology Campus Ghent, Aalst. Her research expertise is in energy performance and indoor climate; interaction building and HVAC systems; commissioning and monitoring for sustainable building construction and has previously worked with KAHO Sint-Lieven, Ghent University and Technum NV in Belgium. Hilde Breesch has published over 45 research papers in leading international journals and conference proceedings and has coordinated several guidelines for industry in the HVAC domain.

Contributors

Karen Allacker Faculty of Engineering Science, Department of Architecture, KU Leuven, Louvain, Belgium Diego Arcos-Aviles Departamento de Eléctrica y Electrónica, Universidad de las Fuerzas Armadas ESPE, Sangolquí, Ecuador Paúl Ayala Departamento de Eléctrica y Electrónica, Universidad de las Fuerzas Armadas ESPE, Sangolquí, Ecuador Hilde Breesch Department of Civil Engineering, KU Leuven, Construction Technology Cluster, Technology Campus Ghent, Sustainable Building, Ghent, Belgium Muhammad Burhan Water Desalination and Reuse Centre (WDRC), Biological and Environmental Science & Engineering (BESE), King Abdullah University of Science and Technology, Thuwal, Saudi Arabia Silvia Coccolo Solar Energy and Building Physics Laboratory (LESO-PB), Ecole Polytechnique Fédérale de Lausanne (EPFL), Lausanne, Switzerland Frank De Troyer Faculty of Engineering Science, Department of Architecture, KU Leuven, Louvain, Belgium Ngo Hoang Ngoc Dung National University of Civil Engineering, Hanoi, Vietnam Pietro Florio Solar Energy and Building Physics Laboratory (LESO-PB), Ecole Polytechnique Fédérale de Lausanne (EPFL), Lausanne, Switzerland Editors and Contributors xiii

Rodolfo Gordillo Departamento de Eléctrica y Electrónica, Universidad de las Fuerzas Armadas ESPE, Sangolquí, Ecuador Francesc Guinjoan Department of Electronics Engineering, Escuela Técnica Superior de Ingenieros de Telecomunicación de Barcelona, Universitat Politècnica de Catalunya, Barcelona, Spain Nguyen Trung Kien Vilandco Company. Hanoi, Hanoi, Vietnam Martin P. Marietta Department of Electronics Engineering, Escuela Técnica Superior de Ingenieros de Telecomunicación de Barcelona, Universitat Politècnica de Catalunya, Barcelona, Spain Luis Marroyo Department of Electrical and Electronics Engineering, Public University of Navarre (UPNa) Ediﬁcio de los Pinos, Pamplona, Spain Carmen-Pilar Martí-Ballester Universitat Autònoma de Barcelona, Bellaterra, Spain Dasaraden Mauree Solar Energy and Building Physics Laboratory (LESO-PB), Ecole Polytechnique Fédérale de Lausanne (EPFL), Lausanne, Switzerland Emilia Motoasca Faculty of Engineering Technology, Department of Electrical Engineering, KU Leuven Technology Campus Ghent, Ghent, Belgium Vahid M. Nik Division of Building Physics, Department of Building and Environmental Technology, Lund University, Lund, Sweden; Division of Building Technology, Department of Civil and Environmental Engineering, Chalmers University of Technology, Gothenburg, Sweden; Institute for Future Environments, Queensland University of Technology, Garden Point Campus, Brisbane, QLD, Australia Kim Choon Ng Water Desalination and Reuse Centre (WDRC), Biological and Environmental Science & Engineering (BESE), King Abdullah University of Science and Technology, Thuwal, Saudi Arabia Quang Minh Nguyen National University of Civil Engineering, Hanoi, Vietnam Adedayo J. Ogungbile Department of Quantity Surveying, School of Environmental Technology, The Federal University of Technology, Akure, Nigeria Ayodeji E. Oke Faculty of Engineering and Built Environment, Department of Construction Management and Quantity Surveying, University of Johannesburg, Johannesburg, South Africa Wout Parys Building Physics Section, Department of Civil Engineering, KU Leuven, Heverlee, Belgium Julio Pascual Department of Electrical and Electronics Engineering, Public University of Navarre (UPNa) Ediﬁcio de los Pinos, Pamplona, Spain xiv Editors and Contributors

A. T. D. Perera Solar Energy and Building Physics Laboratory (LESO-PB), Ecole Polytechnique Fédérale de Lausanne (EPFL), Lausanne, Switzerland J. Peuteman M-Group (Mechatronics), KU Leuven, Campus Bruges, Bruges, Belgium Delphine Ramon Faculty of Engineering Science, Department of Architecture, KU Leuven, Louvain, Belgium Dirk Saelens Building Physics Section, Department of Civil Engineering, KU Leuven, Heverlee, Belgium; EnergyVille, Genk, Belgium Pablo Sanchis Department of Electrical and Electronics Engineering, Public University of Navarre (UPNa) Ediﬁcio de los Pinos, Pamplona, Spain Jean-Louis Scartezzini Solar Energy and Building Physics Laboratory (LESO-PB), Ecole Polytechnique Fédérale de Lausanne (EPFL), Lausanne, Switzerland Muhammad Wakil Shahzad Water Desalination and Reuse Centre (WDRC), Biological and Environmental Science & Engineering (BESE), King Abdullah University of Science and Technology, Thuwal, Saudi Arabia Alvin G. Stern AG STERN, LLC, Newton, MA, USA Damien Trigaux Department of Architecture, KU Leuven, Louvain, Belgium; EnergyVille, Genk, Belgium; VITO, Unit Smart Energy and Built Environment, Mol, Belgium Nicole P. M. van Lipzig Faculty of Science, Department of Earth and Environmental Sciences, KU Leuven, Louvain, Belgium Barbara Wauman Department of Civil Engineering, KU Leuven, Construction Technology Cluster, Technology Campus Ghent, Sustainable Building, Ghent, Belgium Hendrik Wouters Faculty of Bioscience Engineering, Department of Forest and Water Management, UGent, Ghent, Belgium Introduction to Energy Sustainability

Abstract The European Union commits itself to develop a sustainable, competitive, secure and decarbonized energy system by 2050. This is not only a European but also a worldwide challenge included in the 17 Sustainable Development Goals of UN, like Goal 7 Ensure access to affordable, reliable, sustainable and modern energy for all, Goal 9 Build resilient infrastructure, promote sustainable indus- trialization and foster innovation, Goal 11 Make cities inclusive, safe, resilient and sustainable, Goal 12 Ensure sustainable consumption and production patterns and even Goal 13 Take urgent action to combat climate change and its impacts. Energy sustainability is more than restricting the energy use in industry, buildings and systems or the simple use of renewable energy sources. This calls for a multidisciplinary approach in various economic domains and at various scales. This book provides a holistic approach in terms of energy sustainability implementation, technologies and strategies.

Keywords Energy efﬁciency • Sustainable energy production and storage Neighbourhood and buildings

Meeting the energy efficiency (and sometimes the extra energy self-sufficiency) criteria is still a worldwide challenge in all applications (industry, buildings, consumer goods and services, etc.). Energy sustainability does not mean only reducing the energy use and using more energy produced from renewable energy sources of increasing the energy efficiency of processes and devices, but it implies insight and knowledge in various domains. The publication of this book has been motivated by an increased interaction among various disciplines (electrical, mechanical, construction and material engineering, economics, politics, etc.) in various economic domains (industry, constructions, etc.) and at various scales (worldwide, nation- wide, neighbourhoods, buildings, devices) where newest developments in solar, wind and waste heat energy harvesting, hydrogen production and energy storage technologies are applied. Besides the scale level, each country deals with own geographic, economic, politic and legal particularities that influence the way some of the energy sustainability challenges are approached and coped with. Therefore,

xv xvi Introduction to Energy Sustainability this book has a broad view and tackles a wide range of topics including energy sustainability implementation, technologies and strategies. Meanwhile, this book shows just a small selection of the multidisciplinary field of energy sustainability. In the first four chapters, bundled as Part I—Energy Sustainability Implementations, some aspects of the implementation at macro-scale (nation, city, neighbourhood, etc.) and at building scale related to energy sustainability are presented. Chapter 1 investigates the sustainable construction practices in West African countries in the context of recent challenges, drivers and implementation and application measures. The increasing interest in sustainable buildings has been driven by a strong increase among relevant stakeholders including clients, sponsors, construction professionals, government agencies and other concerned regulatory bodies. Chapter 2 studies the influence of urban planning in Belgium on the financial and environmental impact of neighbourhoods using a combined life-cycle costing (LCC) and environmental life-cycle assessment (E-LCA). The results reveal substantial impact differences between different neighbourhoods, showing the importance of urban planning to decrease the financial and environmental impact of the built environment. Chapter 3 makes a detailed SWOT analysis together with an extended discussion on the application of high-tech and low-tech designs for reaching the energy efficiency in residential buildings in Vietnam. Up-to-date carefully chosen data and examples are provided. Chapter 4 further discusses the specific situation of Vietnam, providing an overview of the actual design practice for office buildings in Vietnam together with design recommendations for office buildings to achieve high standards of the energy efficiency and indoor climate. In the next five chapters, grouped as Part II—Energy Sustainability Strategies, various energy sustainability strategies are described in relation to buildings, neighbourhoods, systems and energy markets. Chapter 5 focuses on the energy efficiency and sustainability on the urban scale and elaborates how to develop a computational platform combining future climate conditions, assessment of energy demand of a building stock and design and assessment of urban energy systems. Chapter 6 also considers the importance of accounting for climate changes, more specifically in dynamic building energy simulations. This issue is clearly needed when looking for sustainable buildings as buildings have a relatively long lifespan. The chapter discusses widely used methods to predict future weather data and provides an overview of available weather data sets for building simulations. Another essential aspect to achieve energy sustainability in buildings is the application of a reliable and accurate method for the energy use assessment of building designs. Chapter 7 assesses the accuracy of a simplified calculation method in office and school buildings in Belgium by using integrated dynamic building and HVAC system simulations. The simplified approach as currently applied in the EPR calculation tool in Flanders is shown to be suited for the calculation of the final energy use, despite the uncertainties and restrictions of the investigated simulation model. Chapter 8 discusses a fuzzy-based approach to design control strategies for microgrids, where the residential grid-connected microgrids (MGs) that comprise renewable generation and storing capability are constrained to grid operator requirements which include a smooth and bounded grid power profile. Chapter 9 Introduction to Energy Sustainability xvii ends Part II with an analysis of the effectiveness of financial instruments to invest in renewable energy on the Spanish market. The financial performance of various alternative energy mutual funds is compared to conventional market benchmarks. In the last five chapters, grouped as Part III—Energy Sustainability Technologies, various technologies related to sustainable energy production, use and storage are considered. Chapter 10 provides a detailed overview of basic concepts related to wind energy: energy calculations, design methodology, construction and electrical power generation using wind turbines. Chapter 11 presents the newest developments and challenges related to the use of thermoelectric materials for waste heat recovery in various applications. Thermoelectric generators based on thermoelectric materials have the capability of converting heat energy into electric energy and therefore have an immense potential to increase the energy efficiency of various processes and devices. Chapter 12 discusses a sustainable approach for cooling needs using concentrated photovoltaic (CPV) in combination with mechanical vapour compression (MVC) and adsorption chillers. The thermal energy recovered from the cooling of CPV system is used in the absorption chillers, and this leads to a strong increase of the system efficiency. Chapter 13 describes an economical, novel method for implementing a complete hydrogen fuel clean energy cycle based on the chemical reaction between salinated (sea) or desalinated (fresh) water and sodium metal with the use of a novel, molten salt electrolytic cell designed to perform electrolysis at a temperature range between 950 °C and 1050 °C. Chapter 14 concludes Part III with a discussion of the use of renewable energy storage at KAUST desalination plant pilot to increase the solar-driven desalination capacity. The topics are organized into three different sections: (i) energy sustainability implementation, (ii) energy sustainability strategies and (iii) energy sustainability technologies. Specific topics covered in this book include: • Sustainable construction practices in West African countries, • Influence of urban planning on the financial and environmental impact of neighbourhoods, • Energy efficiency in urban residential buildings in Vietnam, • Recommendations for the design of office buildings in Vietnam, • Computational platform linking neighbourhoods to sustainable energy systems, • Future weather data for dynamic building energy simulations, • Simplified method to assess heating energy use in non-residential buildings, • Microgrid energy management strategies, • Financial instruments to invest in renewable energy in Spanish market, • Overview of basic concepts related to wind energy, • Newest developments and challenges related to thermoelectric materials for waste heat recovery, • Concentrated photovoltaic thermal (CPVT) system for cooling, • Novel method for implementing a complete hydrogen fuel clean energy cycle, • Application of renewable energy storage for desalination. xviii Introduction to Energy Sustainability

To summarize, this book contains information about energy sustainability implementation at macro-scale (nation, city, neighbourhood) and building scale, energy sustainability strategies in relation to buildings, neighbourhoods, systems and energy markets and sustainable energy production, use and storage technologies. We sincerely hope that you will enjoy its content!

Emilia Motoasca Avinash Kumar Agarwal Hilde Breesch Part I Energy Sustainability Implementation Chapter 1 Sustainable Construction Practices in West African Countries

Adedayo J. Ogungbile and Ayodeji E. Oke

Abstract The quest for sustainable construction practices has been on the increase among relevant stakeholders including clients, sponsors, construction professionals, government agencies and other concerned regulatory bodies. This article examines the level of practice of sustainable development goals in the West African countries’ construction industry with an emphasis on the challenges, drivers and possible measures for improving its implementation and application. Various gaps and neglected issues in sustainable construction in the region were also reviewed to providing necessary information for the expansion of knowledge of relevant stakeholders and ensuring that construction projects are delivered to international standards.

Keywords Green economy Á Sustainable development Á West Africa

1.1 Introduction

West Africa is a region on the African continent that experiences developmental activities year in, year out and the type of developments encountered in the area can, however, cannot compete with that obtained in other parts of the word (Adebayo and Adebayo 2000). An overview of developmental standard in place in the West African region and that in the advanced world can be said to be one of the reasons why most countries in the area are classiﬁed as developing countries (Du Plessis 2002). One of the Millennium Development Goals, as seen in the Cities

A. J. Ogungbile (&) Department of Quantity Surveying, School of Environmental Technology, The Federal University of Technology, P.M.B. 704, Akure, Nigeria e-mail: [email protected] A. E. Oke Faculty of Engineering and Built Environment, Department of Construction Management and Quantity Surveying, University of Johannesburg, Johannesburg 2028, South Africa e-mail: [email protected]

© Springer Nature Singapore Pte Ltd. 2019 3 E. Motoasca et al. (eds.), Energy Sustainability in Built and Urban Environments, Energy, Environment, and Sustainability, https://doi.org/10.1007/978-981-13-3284-5_1 4 A. J. Ogungbile and A. E. Oke

Without Slums initiative, targeted that by the year 2020, more than 100 million slum dwellers would have been significantly touched for the good by improving their standard of living through the provision of improved sanitation, clean water and primary health care services. In ensuring that these goals are achieved in West Africa, more developmental initiatives are created, and apparently, the built environment sits as the driver of the whole efforts. Du Plessis (2007) averred that the type of built environment put in place coupled with the nature of its creation will in great ways determine the level of achievement that will be realized by the objectives of the developmental initiatives of the MDGs. Sustainable development strategies are unique (Moon 2013). Every country that has embraced sustainable development has strategies for its operation. There is no agreed strategy to sustainable development. Every country fashioned out strategy that fit into the construction industry and prevailing factors surrounding construction operation of the industry. Sustainable development strategy reflects the priorities and concern of each country (Oke and Aigavboa 2017). The strategies are born out of challenges facing construction operations in the countries. The challenges are indentified alongside with measures to militating or overcoming them. Major factors to be considered in sustainable development strategy are reflection of local value, the involvement of all stakeholders in the community and acceptance of citizen with high sense of ownership. The dire need for a sound and an ever-growing development in physical infrastructure and the built environment in developing countries around the world (West Africa inclusive) needs to be addressed in an economical and more socially responsible way than what used to be the case in time past in order to ensure that they stand the test of time (Du Plessis 2007). Du Plessis (2007) claimed that the emerging nature of the built industry in the region creates an excellent opportunity to make interventions on sustainability in developments now. Considering the continued rapid rural–urban migration happening in developing countries and the increased rate of infrastructural development as a result of the MDGs intervention, the need for the introduction of sustainable construction, however, cannot be delayed further.

1.2 Sustainable Construction

Du Plessis (2002) averred that virtually all sectors of the society have continually been pursuing and interpreting sustainable development and sustainability in their comprehensive framework arising from the adoption and formulation of the Agenda 21 as a blueprint for sustainable development in the 1992 Rio Earth summit. The construction industry is not left out in keying to the sustainable trend as obtained in other works of life (Ogungbile and Oke 2015). Bond and Perrett (2012) averred that water, energy, materials, and land are the vast amounts of resources on which the built environment flourishes. Chances of reducing destruction to the environment keep resurfacing from inception to completion of the whole process. Sustainable 1 Sustainable Construction Practices in West African Countries 5 development as often described by many requires a long-term and joint outlook by the society which integrates environmental, economic and social objectives (Dearing 2010). Cook et al. (2012) argued that respect for human values and opportunity are the basis on which organizing framework of sustainable development is offered. A wide range of organizational responses to sustainability is evident (Du Plessis 2007). Du Plessis (2007) adduced that while some organizations’ responses could be seen as being pragmatic, some others respond by seeing sustainable development as a vision. Du Plessis (2007) submitted that meeting human values and needs in different ways constitute innovation and sustainable development is all about innovation. Accommodating sustainable development to processes of reinventing business would not result in the desired output as compared to having a clear line of direction supported by right metrics, management backing and adequate resources. Whichever the preferred approach is, innovation has always emanated from being presented with (or continuous presenting) a credible tactical dilemma which can be fixed by creating utterly new approaches. The innovative response of sustainable construction is particularly characterized by its eco-efficiency nature. The goal of sustainable construction is targeted towards maintaining the natural balance of the environment and at the same time innovating new approaches to construction. In achieving Eco-efficiency, possible ways of eradicating impact right from the origin were targeted as the best response to environmental impact rather than reducing the obvious impact (Dearing 2010). Du Plessis (2007) similarly suggested that assessing sustainability from obligations and cost would give no future to sustainable development. Continuous assessment in this way would only lead to being caught in the commodity trap (Du Plessis 2007). Provision for long-term profitability and growth by sustainable development is needed to become a successful and integral part of business philosophy. However, sustainable construction could have various meanings to the different parties involved in the construction process. These parties are the constructor or the contractor, the clients and the designers. For a client, sustainable construction could mean an improved corporate relationship image with the locals. The positioning of the clients near to the top of the supply chain in construction place them in playing an important role in delivering more sustainable construction. In recent years, forward-looking clients have continuously asked construction companies and design teams for the construction of more sustainable projects (Dearing 2010). Moreover, for clients, sustainability could mean lower life-cycle costs (construction and future maintenance costs) relating to their project when issues of energy efficiency are concerned (Häkkinen and Belloni 2011). A reduced energy bill can be an example of the energy efficiency feature of sustainable construction (Häkkinen and Belloni 2011). Also, it could signify to a client a better relationship with host communities, i.e. remodel the construction process to be more efficient by involving the local communities more in the whole process; making them an integral part of the project decision-making groups. UNRISD (2013) submitted that the involvement of the local communities in the construction process brings to the process massive benefits such as a smoother planning process devoid of rancour 6 A. J. Ogungbile and A. E. Oke and enhanced reputation of the project contributors. According to UNRISD (2013), healthy and pleasant environment, economy and time savings are brought about by having a more robust trust, respect and dialogue between the various participants of the construction team including the clients and the local communities. This, in turn, will improve staff retention and productivity in the construction process as a result of having a productive working environment. To the same client, sustainable construction could also be termed to be a reduction in environmental impact through both construction and operation. However, to a constructor, sustainability could be interpreted to mean various meanings to differ from that of the client. It could mean a way to better satisfy the client by getting the best solution for relating with designers, suppliers and even the client (UNRISD 2012). It could also mean creating a better reputation with the host community of the project and the constituted authority in such an area by enhancing positive relationship through social responsibilities and conflict avoidance mechanism put in place by the contractor. Also, to the constructor, the most important meaning that could be read of sustainable as a reduction of cost and wastage. It could also signify a reduction in legal action, risk reduction and better use of resources efficiently, improved health and safety and better communication. Sustainable construction has been adjudged to have some compelling advantages some of which are improved well-being; valuable operational costs of sustainable construction; productivity of users and occupants due to improved project performance (Häkkinen and Belloni 2011); use of natural resources and reduced environmental impact as a result of reduced emissions leading to a long-term national economic benefits (Du Plessis 2007).

1.3 Challenges and Barriers to Sustainable Construction in West Africa

All over the world, the face and nature of construction industry keep changing from time-to-time. This is as a result of the continued interest in improving the ways things are being done, and construction projects are being procured, executed and delivered to suit the demanding nature of the ever-changing environment. However, the trend faces some salient and compelling challenges to its success story. Some of these challenges are as follows.

1.3.1 Resistance to New Technologies

In the West African construction industry, one of the major hindrances to achieving a sustainable environment regarding adoption of the sustainable practices as applied in other developed regions of the world is resistance to new technologies (Adebayo 1 Sustainable Construction Practices in West African Countries 7 and Adebayo 2000). Sustainable construction is driven by adherence to new technological systems and procedures. These new technologies are alien to the construction system on the ground in the West African, and this is posing a lot of difficulties to the acceptance and of course, the development of sustainability construction environment in the region. Sustainable construction entails an overhaul process change, and it requires new perception to analysing unforeseen costs and possible risks (Häkkinen and Belloni 2011). Learning ways of networking, roles, actors, new tasks and decision-making phases are needed to create new efficient processes for reducing and overcoming these hindrances. Furthermore, Häkkinen and Belloni (2011) stated further that construction clients need to be more educated about the benefits of sustainable construction to enhance their existing knowledge.

1.3.2 Rules of Competition and Tendering Processes

For every process change, there should be a new rule to the game that guides the process. The case is different in the West African construction industry in general. This is perceived in the way construction tendering activities, and procurement is done. Bidding processes still follow the old methods of conventional tendering which is created a limitation to the thriving of sustainable construction in the region.

1.3.3 Lack of Functioning Value Chains

A value chain is a set of activities that a ﬁrm operating in a speciﬁc industry performs to deliver a valuable product or service to the market. In the construction industry, value chain can be the activities that precede construction, done during project execution and also, the post-construction activities expected to be carried out by contractors, consultants, clients and other parties to construction projects to ensure rational acceptability in the market (the built environment). Despite the emerging advocacy for the need for accepting sustainability in the construction industry, the West African region has lagged in the acceptance because of the poor functioning value chain. Until the value chain in the region is improved upon, sustainable construction in the region will continue to be limited in scope and acceptability in the market, and this possesses a massive challenge to the attainment of a sustainable environment in the region.

1.3.4 Possibilities to Apply Integrated Design Team

Sustainable construction is about ideas. It entails more than the usual architectural, structural and aesthetic designs as obtained in the conventional construction as we 8 A. J. Ogungbile and A. E. Oke are used to. It is majorly teamwork as no one can claim to be an island when it comes to sustainable development. The failure of professionals to work in team continues to be a signiﬁcant challenge to the proper establishment of sustainable concept in the African region (Du Plessis 2007). Majority of the designs are imported ideas from different continents of the world which, in most cases, might not be suitable for adoption in the African region. Häkkinen and Belloni (2011) emphasized that sustainable construction ideas require full adherence to details. This full adherence may not be achieved when working with imported designs. Therefore, a need for a holistic review of the nature of collaboration exists between construction professionals in the region.

1.3.5 Lack of Demand

The nature of the complexity involved in sustainable construction naturally makes it more expensive to procure when compared to the conventional construction system (Bond and Perrett 2012). Due to this costly nature and the financial peculiarity of the West African region, there seemed to be a lack of demand on the part of the government who are the largest clienteles of the construction industry. In the UK, only 40% of construction products are owned by the public sector, a contrast of which is obtained in the African region where a considerable percentage of construction activities are owned, financed and managed by the government. This needs to be improved upon to ensure that more interest come from the private sector regarding financing and ownership of construction activities. Until this is achieved, the demand for sustainable construction would still experience low patronage. The challenges faced by sustainable construction in the West African region are inexhaustible. Some others are lack of adequate knowledge on the part of the constructors, inadequate workforce in handling the new technologies, poor sustainable-focused research, poor marketing process of sustainable construction and so many more.

1.4 Drivers and Enablers of Sustainable Construction in West Africa

Adebayo and Adebayo (2000) alleged that even with the importance of sustainable construction and its contribution to the economy, it had received low attention in Africa. In 2009, The United Nations Environment Programme’s (UNEP) vision for Sustainability in the Building and Construction Sector states that SB is an active process where policies and incentives provided by the government support SB and construction practices, where investors, insurance companies, property developers and buyers/tenants of buildings are aware of sustainability considerations and take 1 Sustainable Construction Practices in West African Countries 9 active roles in encouraging SB and construction practice. The trend has continued in recent years to further preach the need for a sustainable development mentality growth in all parts of the world. Some of the divers of this movement are majorly convened in the advantages that could be derived from the implementation of the sustainable development in the construction industry. The following are some of the benefits of sustainable construction as obtained in advance world where sustainability practice is well established: • Water and energy utility cost savings; • Due to greater occupancy and higher rents in sustainable certified buildings, there is increased Net Operating Income (NOI) leading to higher value ratings of the buildings; • Reduction in cost of maintenance; and • There is an enhancement in occupiers’ productivity level because of the controlled environment and improved health of the tenants. In furtherance of the above listed, Bond and Perrett (2012) gave some other key drivers of sustainable construction to include; • Demonstration of commitment to sustainability and environmental stewardship; • Recruitment and retention of key employees by the sustainable development actors (which could consist of the clients, contractors, developers, etc.); • Public relation benefits advocacy regarding societal sensitization, especially for managers, building owners, and developers; • Marketing benefits, especially for developers and building owners; • Well-guided sustainable target setting; • Adequacy of sustainable design methods and knowledge; • Availability of proper procurement method that fits into the sustainable plan; • Sufficient monitoring mechanisms and procedures; and • State-of-the-art management practices after construction. All the above listed can be said of the West African region if sustainable thinking can be well accepted, established and harnessed by the governments of these nations who remain the largest policy maker and client of the individual construction industry in their countries.

1.5 Gaps and Neglected Issues in Sustainable Construction in West Africa

In recent years, the United Nations has been one of the principal champions for sustainable development. With the focus on the achievement of sustainable development outcomes, many UN agencies have researched into some other critical gaps that need to be looked into (Du Plessis 2002). In 2015, a historic accord was signed by more than 150 nations of the world including world giants (United States 10 A. J. Ogungbile and A. E. Oke of America, United Kingdom, China, Germany, France and Russia to mention but a few), in the capital city of France, Paris. The pact did not leave out the African contingents as all African countries were also duly represented in the meeting. The deal was named ‘The Paris Agreement’ and has continued to make wave ever since then. Although the United States led by Donald Trump has since pulled out of the agreement, the pact still maintains a full acceptance among the comity of nations. However how encompassing the deal is, the responsibilities lie in the hands of the individual countries of the world to fully comply with the agreements contained in the pact regarding implementation and to locally adopt it to the nature and tradition of her locals, to fully enjoy the beneﬁts of the treaty. In the West African region, it is essential that some areas be critically looked into to ensure full adoption and compliance with the Paris Pact. Until these areas are thoroughly looked into, and necessary machinery are put in place, sustainable construction may remain a thing too difﬁcult to achieve in the region. Some of the areas of the gap and neglects include the following.

1.5.1 Security of Lives and Properties

The sense of freedom from fear and want is defined as Human security. It is characterized as ‘the protection to the vital core of all human lives in ways that enhance human freedoms and human fulfilment’ (UNRISD 2013). A sense of protection from danger is a critical tool in sustainable development, and this has for a while, being the advocacy of the UN in preaching for peace all over the world. In the West African region, the depth of human insecurity has been shown by the ongoing and recent multiple crises that have characterized the area in the international news space. Häkkinen and Belloni (2011) testified that this perceived lack of security reveals the extant inadequacy of structural and systemic reforms which has made an equitable development pathway and a socially inclusive environment challenging to achieve in the region. UNRISD (2013) reported that in the past six decades thereabout, significant and substantial progress is being made in many parts of the world in reducing poverty and hardship. Also, it was written in the report that some parts of the world are however left far behind of this substantial progress owing to the presence of con- flicts and repeated violent cycles leading to stagnated social indicators and com- promise in economic developments. This scenario is being played out in many of West African nations as the lives and properties of citizens are in continued danger. The implication of this to sustainability is of the grave and strong impediments to the attraction of development to the region. A release by the World Bank in 2013 stated that over 1.5 billion people, almost a quarter of the total population of the world, inhabit in parts affected by violence, fragility and conflict, the majority of whom are located in the Asian and African continents (UNRISD 2013). Security concerns associated with these volatile areas have been proven to have large bearings on developmental activities in the regions 1 Sustainable Construction Practices in West African Countries 11

(Dearing 2010). Expatriates are not encouraged to invest in the part because of the expensive nature of sustainable construction; it becomes highly not economically advisable to invest in an area of high-security tension and instability. Although, according to Adebayo and Adebayo (2000), there exists a substantial ignorance in the interrelation between development, justice and security in the approach of MDGs (which was claimed to be narrow in scope), the need to reposition human security as a focal of developmental strategies is increasingly gaining recognition. Progress at attaining SDGs is most likely to lag severely without the attention and ensuring security of lives and properties (Adebayo and Adebayo 2000). In furtherance to this, the issue of security is becoming as important as environmental sustainability, social inclusion and economic development in the post-2015 development framework as the concept of sustainable development are being broadened to include more variables (UNRISD 2013). At the centre stage of social policy lies human security. They help in combating social exclusion and in reinforcing social cohesion as more is put into investing and preserving human capital. Du Plessis (2002) held that a number of means could be provided in addressing issues that concern security through various measures of social protection instruments and social policies. This can also be achieved with the enforcement of law and order, ensuring political stability as well as regulation of the labour market. For a country to keep coherent national responses to natural disasters, crises and shocks, as well as ensuring adequate monitoring and reduction of poverty and inequality levels, social security systems are critical Du Plessis (2002). The practice of social security, respect for lives and properties are generally lagging as government apparatus are incapable of achieving these lofty heights as obtained in other parts of the world. For sustainable development to be the order of the day in the West African region, it is necessary that the security of lives and properties of her inhabitants, visitors and expatriates working in the area is well guaranteed to attract sustainable new technologies and investments.

1.5.2 Culture

Culture, considered as ‘the set of distinctive spiritual, material, intellectual and emotional features of a society or a social group’. Culture can be viewed in a broad sense as a critical reflection in defining the constitutive elements of dignity, well-being and sustainable development (Hayashi et al. 2012). Young (1999) acknowledges the exclusion of culture from the MDGs. However, culture as a significant role as it is critical to environmental, economic and social impacts of sustainable development (Young 1999). There existed a missing agreement and shared recognition of culture in the core developmental strategies at global, regional and even local levels in spite of the increasing awareness of the significance of culture in development (UNRISD 2013). A study done by UNESCO in 2010 showed that cultural assets are one of the fastest growing sectors of many economies Hayashi et al. (2012), which underlines the importance of culture to the 12 A. J. Ogungbile and A. E. Oke growth of any economy. In developing countries, like West Africa, where there is abundance substantial labour force, cultural and natural resources, sustainable tourism, creative industries as well as cultural infrastructures and heritage can be pivotal to generating revenues. Culture is held esteemed as it ensures the acknowledgement of indigenous people and minorities, most marginalized groups, in creating a conducive, resilient, stable and more inclusive society (Hayashi et al. 2012). Hayashi et al. (2012) averred that promoting support and respect for cultural expressions can pave the way to consolidating the social capital of a community and foster confidence in public establishments. In promoting more productive patterns and sustainable consumption, traditional and cultural activities are very vital. This can be broadly deployed in tackling ecological challenges. Cultural heritage rehabilitation has been used in various post-conflict situations in helping concerned communities rebuild their lost iden- tities and reclaim their lost common interests. Culture is germane in ensuring sustainable and inclusive development by making proactive policies that institu- tionalize cultural heritage and promote intercultural dialogue. Sustainable development according to UNRISD (2013) should be deeply founded on the involvements of participation of marginalized groups and knowledge of local context. Interpretively, the way of life and the beliefs of the locals should form the basis of any development strategies if it is going to be sustainable. Development or construction that fails to recognize, promote and ensure the continued protection of the cultural heritage of the people has failed the sustainability test even before inception. The cumulative experiences of countries where cultural diversity is a fundamental attribute of society have shown a wide array of policy approaches and all of them represent instances where accommodation of diversity has been a central aspect of government. Apparently, African countries especially the ones in the West have been recognized on the world stage to be rich in culture and traditions. The government of these nations have got a lot to do to preserve her dynamic culture by incorporating any sustainable development plan they have not just to fit into the existing culture of their people but also to complements them. By so doing, sustainable construction and development as a whole would be accepted by the people who are essential drivers of any sustainable endeavour.

1.5.3 Technology

In the sustainable development agenda plans, Moon (2013) attested that there had been increased advocacy for technology as it is an important element in achieving its goals. The term technology includes information and communication technology (ICT) and the process of technology transfer from one sector to another. The global economy is observed to be deepened through the integration of the rapid change in technology, most especially, in ICTs. This trend has provided a means of 1 Sustainable Construction Practices in West African Countries 13 decoupling resource use from growth and creating a new opportunity for the developing countries to join the international production network (SDSN 2013). As to this end, the vital role of technology in the transition process to sustainable consumption and production cannot be overemphasized. In advanced counties of the World, technological innovations have been deployed to improving primary infrastructural development, education, healthcare and general public service deliveries. Even in migration processes, technology has been consistently used to keep things in check. Migrants are enabled to maintain ties with their homelands and making it easy for people to migrate with greater frequency over longer distances (UNRISD 2013). Also, in new media, in reducing corruption, broadening participation, ensuring accountability of public institutions, technology has been helpful (Moon 2013). The importance of technology, especially ICT to all aspects of life in this changing world cannot be underplayed. In spite of the growing trend and contribution of ICT, the African continent as a whole has struggled to meet up with the dynamic movement of the technological world. This failure to meet up with the ICT advancements is a problem to the achievement of sustainable development in the region. In improving access to technologically provided opportunities, it is important to make sure that technological changes are directed to more equitable and sustainable outcomes, particularly in developing countries such as the West African region. It is important to note that the construction industry of the nations of the West African region needs to embrace the technological advances obtained in developed countries by making room for ﬂexibilities in design, execution and monitoring of construction projects.

1.5.4 Green and Fair Economy

In recent years, green economy to a large extent has been at the centre of more of the contemporary efforts to promote sustainable development. Fairhead et al. (2012) opined that a lot of issues had been highlighted by viewing the green economy through a social lens that is infrequently recognized in policy circles. Studies have shown that a win-win assumption about the green economy has taken attention and of which have continued to gather criticism as to its reliability. This has continued to raise serious questioning (Fairhead et al. 2012). It is however suggested that strategies impact social groups and green economy initiatives differently can result in winners and losers’ situation. Schemes and incentives associated with payments for environmental services (PES), monetary pricing and market-based allocation of ecological assets and biofuels often beneﬁt or target the better-off, redistribute assets upwards and favour people and places with the highest purchasing power (UNRISD 2012). Cook et al. (2012) averred that green grabbing which is an extension of land grabbing, a scenario where natural and resources are appropriated for environmental ends. Such ﬁndings suggest that importance must be placed on issues of the green and fair economy to ensure an economic transition that curtails tensions between the 14 A. J. Ogungbile and A. E. Oke environment and economic development. In achieving this, social drivers associated with the community-based development and social policies are put in place. This will ensure more equitability by the government to the people directly affected by the life impacts of the project. In this century, people/host communities have their voices heard. Whereas in this part of the world, the same cannot be said of that feet. For construction to be sustainable, the locals must give the go-ahead before the project commences as they directly involved in the milestone decision-making for the community.

1.6 Conclusion and Recommendation

Sustainable construction is one of the signiﬁcant topical trends in world discuss that is fast becoming a way of life in most of the advanced countries of the world. However, same cannot be said of the West Africa region, and with extension, the African continent as a whole as construction still follows the conventional methods. Apparently, for the region to gain international recognition it deserves and she craves for, efforts need to be made towards full adoption of sustainable development goals in the region and the African continent as an extension.

References

Adebayo AA, Adebayo P (2000) Sustainable housing policy and practice-reducing constraints and expanding horizons within housing delivery. In: Paper presented in 2nd South African Conference on sustainable development in the built environment, Pretoria, South Africa Bond S, Perrett G (2012) The key drivers and barriers to the sustainable development of commercial property in New Zealand. J Sustain Real Estate 48–77 Cook S, Utting P, Smith K (2012) Green economy or green society? In: Contestation and policies for a fair transition, occasional paper 10 Dearing A (2010) Sustainable innovation: drivers and barriers. United Nations, Geneva Du Plessis C (2002) Agenda 21 for sustainable construction in developing countries. In: Pretoria: a discussion document, report for CIB and UNEP–IETC (2002) Du Plessis C (2007) Agenda 21 for sustainable construction in developing countries. CSIR, Pretoria Fairhead J, Leach M, Scoones I (2012) Green grabbing: a new appropriation of nature? J Peasant Stud 39(2):237–261 Häkkinen T, Belloni K (2011) Barriers and drivers for sustainable building. Build Res Inf 39 (3):239–255 Hayashi N, Boccardi G, Hassan NA (2012) Culture in the post-2015 sustainable development agenda: why culture is key to sustainable development. Background note for UNESCO’s high-level discussion on culture in the post-2015 sustainable development agenda Moon B (2013) For the role of technology in sustainable development see United Nations technical support team issues brief conceptual issues. In: New York: report of the secretary-general science, technology and innovation, and the potential of culture, for promoting sustainable development and achieving the millennium development goals for ECOSOC’s 2013 annual ministerial review 1 Sustainable Construction Practices in West African Countries 15

Ogungbile AJ, Oke AE (2015) Sustainable facility management practices in public buildings in Nigeria. In: Ogunsemi DR, Awodele OA, Oke A (eds) Conﬂuence of research, theory and practice in quantity surveying profession for a sustainable built environment. The Nigerian Institute of Quantity Surveyors, Akure, pp 830–843 Oke AE, Aigavboa CO (2017) Sustainable value management for construction projects. Springer, Switzerland SDSN (2013) An action agenda for sustainable development. In: New York: report for the secretary-general, prepared by the leadership council of the sustainable development solutions network UNRISD (2013) Emerging issues: social drivers of sustainable development. United Nations Research Institute for Social Development, New York UNRISD (2012) Social dimensions of green economy. Res Policy Brief 12 Young C (1999) The accommodation of cultural diversity: case studies. Palgrave Macmillan, UNIRSD Chapter 2 Modelling the Inﬂuence of Urban Planning on the Financial and Environmental Impact of Neighbourhoods

Damien Trigaux , Karen Allacker and Frank De Troyer

Abstract Urban planning decisions related to the urban form, built density and neighbourhood location may affect the sustainability of neighbourhoods to an important extent. This chapter investigates the influence of urban planning on the financial and environmental impact of neighbourhoods. A number of schematic neighbourhood models with various layouts and built densities are analysed using an integrated life cycle approach, combining Life Cycle Costing (LCC) and Environmental Life Cycle Assessment (E-LCA). Furthermore, the influence of the neighbourhood location is assessed by comparing the impact of a rural and urban location. The results reveal substantial impact differences (up to 20–25%) between the neighbourhoods, showing the importance of good urban planning to decrease the financial and environmental impact of the built environment. The main reasons for these variations are the lower primary land use, lower energy use for heating and lower material use in high built-density neighbourhoods and compact buildings. Also, the neighbourhood location proved to be a key parameter to decrease the impact of user transport in neighbourhoods, with impact reductions up to 25–30% in an urban area.

Keywords Integrated life cycle approach Á Life cycle costing Environmental life cycle assessment Á Neighbourhood layout Á Building design

D. Trigaux (&) Á K. Allacker Á F. De Troyer Department of Architecture, KU Leuven, Kasteelpark Arenberg 1/2431, 3001 Louvain, Belgium e-mail: [email protected] D. Trigaux EnergyVille, Thor Park 8310, 3600 Genk, Belgium D. Trigaux VITO, Unit Smart Energy and Built Environment, Boeretang 200, 2400 Mol, Belgium

© Springer Nature Singapore Pte Ltd. 2019 17 E. Motoasca et al. (eds.), Energy Sustainability in Built and Urban Environments, Energy, Environment, and Sustainability, https://doi.org/10.1007/978-981-13-3284-5_2 18 D. Trigaux et al.

2.1 Introduction

The higher scale of the built environment has become an important focus in sustainable decision-making. Urban sprawl, which characterizes the Belgian building stock, has a major impact on required infrastructure, energy use, land use and transport (European Environment Agency 2006). In order to move towards a sustainable built environment, neighbourhoods need to be planned differently, focussing not only on the characteristics of individual buildings but also on the urban layout, built density and relations between buildings and their surroundings. Various studies focused on the urban morphology and its impact on energy consumption in buildings, considering aspects such as building compactness, solar gains, access to daylight and natural ventilation (Ratti et al. 2005; Salat 2009). Compared to these studies, this chapter analyses the influence of urban planning in a wider sustainability perspective, including both the life cycle financial and environmental impacts of neighbourhoods. In addition to operational energy use, impacts related to material use, operational water use, neighbourhood land use and user transport are considered. More specifically, the focus of this research is on the assessment of newly built residential neighbourhoods in the Belgian context. Effects of urban planning decisions related to the urban form, built density and the neighbourhood location are analysed. To investigate these effects, a number of schematic neighbourhood models with various layouts and built densities are assessed, using an integrated life cycle approach, combining Life Cycle Costing (LCC) and Environmental Life Cycle Assessment (E-LCA). Furthermore, a comparison is made between a rural and urban location to gain insight in the influence of the neighbourhood location. The methods and results presented in this chapter are part of a doctoral research on the elaboration of a sustainability assessment method for neighbourhoods (Trigaux 2017). The methods are described in Sect. 2.2 and the results of the analysed neighbourhood models are discussed in Sect. 2.3. Conclusions are for- mulated in the final section.

2.2 Methods

2.2.1 Integrated Life Cycle Approach

The ﬁnancial and environmental impact of the neighbourhood models are assessed based on an integrated life cycle approach, combining Life Cycle Costing (LCC) and Environmental Life Cycle Assessment (E-LCA). This integrated approach was originally developed in the context of the SuFiQuaD research project (‘Sustainability, Financial and Quality evaluation of Dwelling Types’) (Allacker 2010; Allacker et al. 2013a) and has been extended from the building to the neighbourhood scale level (Trigaux 2017). 2 Modelling the Inﬂuence of Urban Planning on the Financial … 19

The ﬁnancial and environmental impacts are assessed over the entire neighbourhood life cycle, considering a lifespan of 60 years (Trigaux 2017). In accordance with the European standards related to the sustainability of construction works (CEN 2010, 2011, 2015), the life cycle of a neighbourhood is divided into three main stages: the before use stage, use stage and End-Of-Life (EOL) stage. The before use stage covers all processes prior to the use of the neighbourhood, i.e. pre-construction (including land purchase and transformation), production of building materials (including raw material extraction, transport to the manufacture and manufacturing), transport to the construction site and construction activities. The use stage includes processes related to maintenance, replacement of building components, operational energy and water use, land occupation and user transport. Finally, the EOL stage covers the demolition activities, waste transport, waste processing and disposal. A detailed description of the life cycle stages and system boundaries can be found in Trigaux (2017).

2.2.1.1 Life Cycle Costing (LCC)

The financial costs during the various life cycle stages are considered in the LCC approach. These include the investment cost, the cost during the use stage and the cost at the end of life. A detailed description of the LCC method used can be found in Trigaux (2017), Allacker (2010), and Allacker et al. (2013a). The financial data are collected from various sources. The cost of building elements is mainly based on the Belgian database ASPEN (2015a, b), combined with product specific data. For the neighbourhood infrastructure, the British Spon’s Price Books (Spon press 2015a, b) are used as a Belgian cost database is lacking. Energy, water and building land prices are based on Belgian statistical data (CREG 2015; Belgian Federal Government 2017a, b; VMM 2015). Concerning the financial cost of user transport, a study of Transport and Mobility Leuven is used, including the calculation of consumer prices for different transport modes (Delhaye et al. 2010, 2017). The life cycle financial cost is calculated as the sum of the present values (for the reference year 2015) of all costs occurring during the neighbourhood life cycle. The economic parameters—in real terms—are based on Belgian statistical data and are summarized in Table 2.1 (left column).

2.2.1.2 Environmental Life Cycle Assessment (E-LCA)

The environmental impacts generated during the whole lifespan of the neighbourhood are assessed using the E-LCA approach. The assessment is based on the E-LCA method developed within the MMG research project (‘Environmental proﬁle of building elements’), commissioned by the Public Waste Agency of Flanders (Allacker et al. 2013b; De Nocker and Debacker 2015). The MMG method is an update of the E-LCA method developed in the SuFiQuaD project, in 20 D. Trigaux et al.

Table 2.1 Economic parameters applied for the financial and environmental costs (real rates above the inflation) (Trigaux 2017) Financial costs (%) Environmental costs (%) Real discount rate 2 0 Real growth rate material and water 0 0 Real growth rate labour 1 – Real growth rate energy 2 0 order to be in line with recent E-LCA standards and guidelines in Europe (CEN 2011, 2013; EC-JRC 2011). Concerning the Life Cycle Inventory (LCI), the Swiss Ecoinvent database (version 2.2) is used to collect the input–output flows related to the building materials and processes which are required for the environmental impact assessment (Frischknecht et al. 2007). Preference is given to Western European processes to ensure the representativeness for the Belgian context. When generic Western European processes are lacking, Swiss data records are adapted by replacing the Swiss electricity mix and transport processes by corresponding European processes (Allacker et al. 2013b). Regarding the selected environmental indicators, the impact categories in the MMG E-LCA method include the ones defined by the EN 15804+A1 standard (CEN 2013), which are further referred to as CEN indicators (Table 2.2). In addition, seven more impact categories are considered based on the International Reference Life Cycle Data System (ILCD) Handbook (EC-JRC 2011). The additional impact categories are further referred to as CEN+ indicators (Table 2.3). In addition to the individual environmental impact indicators, the MMG method provides an aggregated single-score indicator, expressed in a monetary value (EURO). This aggregated score indicates the external environmental cost, i.e. the cost to avoid, reduce or compensate the damage caused by environmental impacts to a given level considered to be sustainable. The environmental cost is calculated by multiplying the environmental impact indicator values with their specific monetary value and adding these up to obtain the overall environmental cost (single score). The background for the determination of the monetary values is described in (De Nocker and Debacker 2015). In this research, the MMG monetary values of the central scenario for Western Europe are selected for the E-LCA calculations (Tables 2.2 and 2.3). Compared to other weighting methods, the advantage of expressing environmental impacts in monetary values is the possibility to internalize environmental externalities by calculating the sum of the financial and environmental costs. In contradiction to the financial cost calculations, no discounting of future environmental costs is applied. A real growth rate and discount rate of 0% are used so that future environmental costs are equally valued as present environmental costs (see Table 2.1—right column). This approach is chosen in order to avoid burden shifting in time (Trigaux 2017). 2 Modelling the Influence of Urban Planning on the Financial … 21

Table 2.2 CEN indicators and monetary values (central scenario) (De Nocker and Debacker 2015) CEN indicators Unit Monetary value (€/unit)

Global warming kg CO2 equiv. 0.1 Ozone depletion kg CFC-11 equiv. 49.1

Acidiﬁcation of soil and water kg SO2 equiv. 0.43

Eutrophication kg (PO4) 3-equiv. 20 Photochemical ozone creation kg ethene equiv. 0.48 Depletion of abiotic resources—elements kg Sb equiv. 1.56 Depletion of abiotic resources—fossil fuels MJ, net caloric value 0

Table 2.3 CEN+ indicators and monetary values (central scenario) (De Nocker and Debacker 2015) CEN+ indicators Unit Monetary value (€/unit) Human toxicity—cancer effects CTUh 665,109 Human toxicity—non-cancer effects CTUh 144,081 Particulate matter kg PM2,5 equiv. 34 Ionizing radiation—human health kg U235 equiv. 9.7E−04 Ionizing radiation—ecosystems CTUe (per kBq) 3.7E−05 Ecotoxicity—freshwater CTUe 3.7E−05 Water scarcity m3 water equiv. 0.067 Land occupation—soil organic matter kg C deﬁcit 2.7E−06 Land occupation—biodiversity Urban m2a 0.3 Agricultural m2a 6.0E−03 Forest m2a 2.2E−04 Land transformation—soil organic matter kg C deﬁcit 2.7E−06 Land transformation—biodiversity m2 Not available

2.2.2 Element Method for Cost Control

Due to the complexity of neighbourhoods, a well-structured evaluation is required to deal with the huge amount of data. In the SuFiQuaD method, the assessment structure is based on the element method for cost control (Allacker 2010). The basic principle is the hierarchical subdivision of buildings into building elements, such as external walls, roofs and technical services. These building elements can then again be subdivided into several work sections which are composed of one or more building materials (Fig. 2.1). In consequence, an analysis can be made at various scale levels, each higher level building on the lower levels. The following levels are considered: building 22 D. Trigaux et al.

Fig. 2.1 Element method for cost control and scale levels (Trigaux et al. 2014) materials (e.g. brick, mortar, plaster), work sections (e.g. brickwork, plasterwork), building elements (e.g. external wall including finishes) and buildings. As defined by Trigaux et al. (2014, 2017), the element method can be extended to evaluate neighbourhoods, which are defined as a combination of buildings, networks (e.g. roads, utilities) and open spaces (e.g. squares, parks and gardens) (Fig. 2.1).

2.2.3 Assessment of Neighbourhood Impact Drivers

The integrated life cycle approach developed by Trigaux (2017) allows to report the neighbourhood life cycle impacts according to ﬁve drivers: material use, operational energy use, operational water use, primary land use and user transport. These drivers are brieﬂy described in the subsequent paragraphs. A detailed description can be found in Trigaux (2017).

2.2.3.1 Material Use

Material use refers to the flows of the construction products during the entire life cycle of the neighbourhood. This includes all the flows related to the production, transport, construction, maintenance, replacement and EOL treatment of the construction products. Scenarios and assumptions used to model the life cycle of the building elements are reported in the publications of the SuFiQuaD and MMG project (Allacker 2010; Allacker et al. 2013a, b). The scenarios used for the road infrastructure and open spaces are defined in Trigaux (2017).

2.2.3.2 Operational Energy Use

The operational energy use includes the energy use for space heating, domestic hot water, ventilation, lighting and appliances. The energy use for space heating is calculated based on simpliﬁed approach to estimate the heating demand during the 2 Modelling the Inﬂuence of Urban Planning on the Financial … 23

Fig. 2.2 Analysis of the direct and diffuse solar radiation using ray tracing techniques in a SketchUp 3D model (Trigaux et al. 2015) master planning of neighbourhoods. This approach consists of a design tool to optimize solar radiation and heating energy use in neighbourhoods, requiring limited input (Trigaux 2017; Trigaux et al. 2015, 2017b). Using a plugin implemented in the 3D modelling software SketchUp (2017), detailed information on solar obstructions is extracted from a 3D neighbourhood model (Fig. 2.2). This information is then used to assess the heating energy demand based on the dynamic Equivalent Heating Degree Day (dEHDD) method (Trigaux 2017; Trigaux et al. 2014, 2017b). The energy use for domestic hot water and for ventilation is assessed using a combination of two standards, i.e. the Passivhaus Projektierungs Paket (PHPP) and the Flemish standard for Energy Performance Certiﬁcates (EPC) (VEA 2013; Feist et al. 2001). Finally, the energy use for appliances and lighting is based on average household electricity consumption for Belgium (VREG 2017). When a photovoltaic (PV) system is installed, the electricity production is subtracted from the total electricity consumption.

2.2.3.3 Operational Water Use

The operational water use includes tap water consumption and wastewater and rainwater discharge. Regarding rainwater discharge, only the amount of rainwater which is transported to a wastewater treatment plant is considered. No additional impact for wastewater treatment is calculated for rainwater discharged to the storm sewer or on-site inﬁltration system. The tap water consumption and volumes of wastewater and rainwater discharged are assessed based on an existing water calculation tool, which was developed by the Belgian Building Research Institute (BBRI) (Flemish Government 2011). 24 D. Trigaux et al.

2.2.3.4 Primary Land Use

Neighbourhoods are responsible for two types of land use interventions: primary land use, i.e. the neighbourhood spatial footprint and secondary land use, i.e. land use associated with the resource extraction, production, transport and EOL treatment of construction products (Allacker et al. 2014). While the environmental impact of secondary land use is assessed as part of the life cycle assessment of the construction products, the environmental impact of primary land use is calculated separately based on the method deﬁned in Trigaux (2017) and Trigaux et al. (2017a). This method, which considers the footprint of buildings, infrastructure and open spaces, assesses both the impacts of land transformation (at the start of the neighbourhood life cycle) and land occupation (during the neighbourhood lifespan). Concerning the ﬁnancial impact of primary land use, only the initial cost of land purchase is included in the assessment. Land occupation taxes during the neighbourhood life cycle are not assessed due to a lack of statistical data on the cadastral income.

2.2.3.5 User Transport

User transport focuses on the transportation of the inhabitants during the neighbourhood life cycle. This is assessed by defining transport profiles based on statistical data on average transport distances per transport mode, as reported in the Research on Transport Behaviour in Flanders (version 5.1) (Declercq et al. 2016). This study includes transport data for the different areas of the Structure Plan Flanders, which are subdivided according to their spatial structure (from rural to more urban areas). A comparison between the transport profiles for the different areas allows to investigate the influence of the neighbourhood location on user transport.

2.2.4 Schematic Neighbourhood Models

To analyse the impact of the urban form and built density, four schematic neighbourhood models are assessed, consisting of, respectively, detached houses (Model 1), semi-detached houses (Model 2), terraced houses (Model 3) and apartments (Model 4) (Fig. 2.3). A real neighbourhood case study, including a mix of building typologies, is not considered in this research as the objective is to investigate the influence of the neighbourhood layout and built density based on clearly distinctive models. The neighbourhood models are inspired by representative neighbourhoods located in the Belgian municipality of Leuven. The schematic models are defined using the bottom-up approach of Berghauser Pont and Haupt (2010), which dis- tinguishes various levels of aggregation from the building to the district scale. The 2 Modelling the Influence of Urban Planning on the Financial … 25

Fig. 2.3 Schematic neighbourhood models, based on Trigaux (2017) models differ in built density with a Floor Space Index (ratio of the total building floor area to the land area), ranging from 0.20 in Model 1 to 1.21 in Model 4. More specifically, the schematic models consist of rectangular buildings which are composed of one or more housing units with a floor area of 150 m2 and four inhabitants per housing unit. The buildings are organized in urban islands around a central public space. In this research, residential neighbourhoods of approxima- tively 400 dwellings are considered. Furthermore, the roads, footpaths, parking facilities, square, gardens, piped and electrical services are included in the model. A detailed description of the neighbourhood models can be found in (Trigaux 2017). Element buildups are selected which are in line with the current building standards in Belgium (year 2017). The buildings are composed of a solid structure consisting of brick walls, concrete floors and a concrete flat roof. The composition of the building elements is summarized in Table 2.4. Roads and parking facilities consist of an asphalt surface layer, while concrete paving stones are selected for the square and footpaths. The composition of the external elements is summarized in Table 2.5. Regarding the energy performance, the buildings are insulated to fulfil the Flemish Energy Performance of Building (EPB) standards of 2017 (Flemish Government 2017) and thermally improved double glazed windows are used. An improved airtightness of 6 m3/h m2 and a ventilation system C (natural supply and mechanical exhaust) are assumed. For space heating, a condensing gas boiler 26 D. Trigaux et al.

Table 2.4 Composition of the building elements (Trigaux 2017) Building element Composition U-value (W/m2K) Floor on grade Concrete slab 15 cm—PUR board 9 cm—screed mix— 0.24 fired clay tiles Foundation In situ concrete foundation n/a Pile foundation Prefab concrete piles (only in apartment building) n/a External wall Facing brick—PUR board 8 cm—insulating hollow brick 0.23 14 cm—gypsum plaster—acrylic paint Load-bearing Acrylic paint—gypsum plaster—hollow brick 14 cm— n/a internal wall gypsum plaster—acrylic paint Non-load-bearing Acrylic paint—gypsum plaster—hollow brick 9 cm— n/a internal wall gypsum plaster—acrylic paint Party wall Acrylic paint—gypsum plaster—hollow brick 14 cm— n/a stone wool 6 cm—hollow brick 14 cm—gypsum plaster —acrylic paint Storey floor Acrylic paint—gypsum plaster—hollow core concrete n/a slab 12 cm—pressure layer—screed mix—fired clay tiles Party floor Acrylic paint—gypsum plaster—hollow core concrete n/a slab 12 cm—pressure layer—rock wool 3 cm—screed mix—fired clay tiles Stairs Concrete staircase—metal banister n/a Flat roof EPDM—PIR board 10 cm—concrete slope layer— 0.24 pressure layer—hollow core concrete slab 12 cm— gypsum plaster—acrylic paint Window PVC frame with thermal interruption—thermally 1.44 improved double glazing (g-value = 0.61) Internal door MDF frame—plain door n/a Piped services Condensing gas boiler—panel radiators—coupled instant n/a hot water production—ventilation type C—rainwater collection Electrical services Electric cables—elevators (only in apartment building)— n/a PV panels combined with radiators is selected. The overall efficiency of the heating system is 92%. For the hot water production, an instant boiler, coupled to space heating is provided. The overall system efficiency for domestic hot water is 85%. To fulfil the EPB requirements regarding the use of renewable energy (Flemish Government 2017), a photovoltaic (PV) system is installed on the flat roofs. This results in a yearly electricity production of 19 kWh/m2 useful floor area for the models consisting of single-family houses (detached, semi-detached and terraced). Due to the lower roof ratio, the PV electricity production in the apartment model is limited to 8 kWh/m2 useful floor area. Concerning rainwater management, the neighbourhood models are in line with the current Flemish Urban Planning Regulation on Rainwater Management 2 Modelling the Influence of Urban Planning on the Financial … 27

Table 2.5 Composition of the external elements (Trigaux 2017) External Composition element Road Geotextile—crushed gravel sub-base—cement bound crushed gravel base— asphalt Footpath Geotextile—cement bound crushed gravel base—concrete paving stones Parking Geotextile—crushed gravel sub-base—cement bound crushed gravel base— facilities asphalt Square Geotextile—crushed gravel sub-base—cement bound crushed gravel base— concrete paving stones Gardens Grass and hedges Piped services Concrete storm sewer and vitriﬁed clay sanitary sewer—drainage ditches— drinking water and gas pipes (HDPE) Electrical Electric and data cables services

(Flemish Government 2016) and include rainwater collection tanks, separate storm and sanitary sewers and drainage ditches for rainwater inﬁltration and buffering. Concerning the assessment of primary land use, the neighbourhood models are assumed to be built on forest land. As less than 80% of the total neighbourhood area is considered to be sealed in all four neighbourhood models, the land use type ‘urban discontinuously built’ is selected to characterize the land use of the buildings, road infrastructure and open spaces. For the building land price, the Flemish average is assumed (187.79 €/m2—excluding taxes). Furthermore, the impact of the neighbourhood location is investigated by additionally comparing the terraced house model located in an urban area and the detached house model located in a rural area. For the urban area, a location in the city of Leuven is assumed. The transport proﬁle ‘Regional urban area—central municipalities’ (Table 2.6) is selected to characterize the transport of the inhabitants. Furthermore, the average building land price for Leuven (282.63 €/m2)is assumed instead of the Flemish average (187.79 €/m2). For the rural area, a location

Table 2.6 Average transport distances for the urban and rural area (Declercq et al. 2016) Transport mode Regional urban area—central Rural area municipalities (km/person/day) (km/person/day) Car 24.80 34.99 Bus 0.33 1.03 Tram/metro 0.00 0.01 Train 2.32 2.12 Bicycle 2.24 1.21 Electric bicycle 0.02 0.10 On foot 0.41 0.61 28 D. Trigaux et al. in Rotselaar, a municipality near Leuven, is considered. The related transport proﬁle ‘Rural area’ (Table 2.6) and building land price for Rotselaar (187.28 €/m2) are assumed.

2.3 Results

In this section, the results of the assessment of the schematic neighbourhood models are discussed. In a first step (Sect. 2.3.1), the influence of the urban form and built density is analysed by comparing the neighbourhood models without defining a specific neighbourhood location. Only the impacts of material use, operational energy use, operational water use and primary land use are analysed. The impact of user transport is not considered in this first analysis as it is assumed not to be influenced by the urban form but rather by the neighbourhood location. In a second step (Sect. 2.3.2), the neighbourhood location is altered and the influence of user transport on the life cycle impacts assessed.

2.3.1 Inﬂuence of the Urban Form and Built Density

2.3.1.1 Environmental Impact

The life cycle environmental cost of the schematic neighbourhood models, over 60 years and expressed in euro per m2 useful floor area, is shown in Fig. 2.4. Large variations can be noticed between the models: the life cycle environmental cost of the models consisting of terraced houses and apartments is up to about 20% lower compared to the detached house model. Based on a detailed analysis of all impact contributors (Figs. 2.5, 2.6, 2.7 and 2.8), three main reasons, in descending order of magnitude, are identified for these variations. First, the impact of primary land use decreases significantly (up to 80%) for higher built densities. This is a consequence of the high reduction of the land use surfaces for gardens and road infrastructure1 in the terraced house and apartment models (Fig. 2.8). Second, compact buildings such as terraced houses and apartments have lower transmission losses through the building envelope, compared to detached houses. This results in a lower energy use for space heating, which is one of the main contributors to the impact of operational energy use (Fig. 2.6). Compared to the detached house model, the environmental cost of operational energy use is about 15% lower in the terraced house model. Third, a higher built

1The contributions to the environmental cost of primary land use are proportional to the land use surfaces as the same types of original land use (forest land) and neighbourhood land use (urban discontinuously built) are assumed for the buildings, road infrastructure and open spaces. 2 Modelling the Inﬂuence of Urban Planning on the Financial … 29

500 Primary land use

450 Operational water use Operational energy use 400 Material use 350

300

250

2 200

150 e u r o / m u s e f u l f l o o r a r e a 100

0 Model 3_terraced Model Model 1_detached Model Model 4_apartment Model Model 2_semidetached Model

Fig. 2.4 Life cycle environmental cost of the neighbourhood models (excluding user transport), subdivided per driver

180 Electrical services (utilities) Piped services (utilities) 160 Gardens 140 Square Parking facilities 120 Footpath 100 Road Electrical services (building) 80 Piped services (building) Internal door 60 Window 40 Flat roof euro/m² useful floor area floor useful euro/m² Stairs 20 Internal floor 0 Internal wall External wall Foundation Floor on grade Model 3_terraced Model Model 1_detached Model Model 4_apartment Model Model 2_semidetached

Fig. 2.5 Environmental cost of material use for the neighbourhood models, subdivided per building and external element 30 D. Trigaux et al.

200 Road lighting 180 Lighting and appliances Ventilation 160 Domestic hot water 140 Heating

120

100 useful floor area floor useful

≤ 80

60 euro/m 40

0 Model 3_terraced Model Model 1_detached Model Model 4_apartment Model Model 2_semidetached Model

Fig. 2.6 Environmental cost of operational energy use for the neighbourhood models, subdivided per energy contributor

45 Rainwater discharge from paved areas Rainwater discharge from roofs 40 Wastewater discharge 35 Drinking water consumption

10 euro/m² useful floor area floor useful euro/m²

0 Model 3_terraced Model Model 1_detached Model Model 4_apartment Model Model 2_semidetached

Fig. 2.7 Environmental cost of operational water use for the neighbourhood models, subdivided per contributor 2 Modelling the Inﬂuence of Urban Planning on the Financial … 31

100 Land use square

90 Land use parking facilities Land use road infrastructure 80 Land use gardens 70 Land use buildings

euro/m² useful floor area floor useful euro/m² 20

0 Model 3_terraced Model Model 1_detached Model Model 4_apartment Model Model 2_semidetached

Fig. 2.8 Environmental cost of primary land use for the neighbourhood models, subdivided per contributor density results mostly in less material use for buildings, networks and open spaces due to lower element ratios (Fig. 2.5). Reductions in material environmental cost of about 10% are noticed between the models consisting of detached and terraced houses. Despite the higher built density, the life cycle environmental cost of the apartment model is similar to the terraced house model (Fig. 2.4). Three reasons are identiﬁed. First, there is a lower potential for PV electricity production in the apartment model due to the lower roof ratio. This results in a higher impact of operational energy use for lighting and appliances for the apartment model compared to the terraced house model (Fig. 2.6). Second, the impact of material use in apartment buildings is similar to the terraced houses due to the additional impact of collective circulation spaces and the high impact of pile foundations (Fig. 2.5). Third, the impact of operational water use is slightly higher for the apartment model due to the lower potential for rainwater reuse from roofs (Fig. 2.7).

2.3.1.2 Financial Impact

When analysing the financial cost, a similar picture is obtained regarding the influence of the urban form and built density (Fig. 2.9). The life cycle financial cost of the terraced house and apartment models is up to about 25% lower compared to the detached house model. As for the environmental cost, the life cycle financial cost of the apartment model is similar to the terraced house model, despite the 32 D. Trigaux et al.

8000 Primary land use Operational water use 7000 Operational energy use Material use 6000

5000

4000

2 3000

2000 e u r o / m u s e f u l f l o o r a r e a

1000

0 Model 3_terraced Model Model 1_detached Model Model 4_apartment Model Model 2_semidetached

Fig. 2.9 Life cycle financial cost of the neighbourhood models (excluding user transport), subdivided per driver higher built density. The main difference between the financial and environmental cost is the contribution of the various drivers. While operational energy use and material use are the main contributors to the environmental cost, the financial cost is dominated by the impact of material use, which contributes to more than 70% of the life cycle cost in all neighbourhood models. A detailed analysis of all contributors to the financial impact of material use, operational energy use, operational water use and primary land use can be found in Trigaux (2017). Similar to the environmental cost, the main reasons for the financial impact variations (in descending order of magnitude) are the lower material use, lower primary land use and lower energy use for heating in high built-density neighbourhoods and compact buildings.

2.3.2 Inﬂuence of the Neighbourhood Location

In this section, the combined effect2 of the built density and neighbourhood location is analysed by comparing the detached house model in a rural area and the terraced

2The built density and neighbourhood location are not fully independent parameters as rural areas are mainly characterized by a low-built density and urban areas by a high-built density. It is therefore chosen to analyse the combined effect of both parameters. 2 Modelling the Inﬂuence of Urban Planning on the Financial … 33

1000 User transport

900 Primary land use Operational water use 800 Operational energy use 700 Material use

600

500

2 400

300 e u r o / m u s e f u l f l o o r a r e a 200

100

0 Detached_rural area Terraced_urban area

Fig. 2.10 Life cycle environmental cost of the detached and terraced house models, located, respectively, in a rural and urban location. The environmental cost is subdivided per main impact driver

12000 User transport Primary land use 10000 Operational water use Operational energy use Material use 8000

6000 2 4000 e u r o / m u s e f u l f l o o r a r e a 2000

0 Detached_rural area Terraced_urban area

Fig. 2.11 Life cycle financial cost of the detached and terraced house models, located, respectively, in a rural and urban location. The financial cost is subdivided per main impact driver house model in an urban area. The life cycle environmental and financial costs (including user transport) of both models are shown in Figs. 2.10 and 2.11. The environmental and financial cost of the terraced house model is, respectively, 25% and 23% lower, compared to the detached house model. The neighbourhood location has a high influence on the impact of user transport. The environmental and financial cost of user transport in an urban area is about 25–30% lower compared to a rural area. This is mainly due to the car transport distances, which are about 30% lower for people living in urban areas (Table 2.6). 34 D. Trigaux et al.

2.4 Conclusions

2.4.1 Main Results and Conclusions

In this research, the influence of urban planning on the financial and environmental impact of newly built residential neighbourhoods is investigated. Four schematic neighbourhood models with various built densities are assessed based on an integrated life cycle approach, combining Life Cycle Costing (LCC) and Environmental Life Cycle Assessment (E-LCA). The proposed approach includes an assessment of the main neighbourhood impact drivers, i.e. material use, operational energy use, operational water use, primary land use and user transport. The hierarchic and modular structure used for modelling the neighbourhoods proves to allow to identify the impact drivers and hence support urban planning decisions. The assessment of the neighbourhood models shows the high influence of the urban form and built density. The life cycle environmental and financial costs (excluding user transport) of the terraced house and apartment models are up to, respectively, 20% and 25% lower compared to the detached house model. Higher built densities and compactness lead to a lower primary land use, lower energy use for heating and lower material use for buildings, networks and open spaces. However, the impact reductions tend to flatten for high built densities as the life cycle impacts of the apartment model are found to be similar to the terraced house model. The comparison between a rural and urban location reveals the major influence of the neighbourhood location. The life cycle financial and environmental impact of user transport is about 25–30% lower in an urban area. Based on these analyses, it can be concluded that good urban planning, focussing on both urban form, built density and neighbourhood location is one of the key parameters to reduce the financial and environmental impact of neighbourhoods.

2.4.2 Further Research

The research presented in this chapter has a number of limitations which should be addressed in further research (Trigaux 2017). The limitations are related to the assessment method, application of the method, result uncertainties and assessment of future changes. Concerning the assessment method, the study is limited to the effects of urban planning on the financial and environmental impacts of neighbourhoods. An assessment of the social impacts and neighbourhood qualities is not included although both aspects are also essential when striving for sustainable neighbourhoods. Furthermore, both local climate impacts and heat island effects have not 2 Modelling the Influence of Urban Planning on the Financial … 35 been considered in this research. Their influence on the operational energy use could be added in further research. Regarding the application of the method, three limitations can be mentioned. First, the scope of the research is limited to newly built residential neighbourhoods. The proposed approach is nevertheless extendable for the assessment of mixed-use neighbourhoods and refurbishment projects. Second, the analysed case studies are schematic neighbourhood models which are selected for their representativeness for the Belgian context. The assessment method nevertheless allows to assess more complex and innovative neighbourhood types. Third, the analysis of the services for space heating and domestic hot water is limited to individual systems. The method should, therefore, be extended to assess collective systems such as district heating and inter-building energy exchanges, as an increasing importance of such systems is noticed in new neighbourhood developments. Concerning the result uncertainties, the life cycle impact calculations are based on a huge number of input data, scenarios and assumptions. To improve the validity of the conclusions, detailed sensitivity analyses should be done considering the economic parameters, the uncertainties related to the environmental and financial data, the implemented life cycle scenarios and the calculation methods used. Finally, the assessment focuses on the current Belgian building practice. However, fundamental changes are expected in future such as the rise of a sharing economy, the evolution towards a more circular economy, a decarbonation of the energy production and a shift in more sustainable transport modes. The effect of these future changes should be investigated as they may influence the results importantly.

References

Allacker K (2010) Sustainable building, the development of an evaluation method. PhD dissertation, KU Leuven Allacker K, De Troyer F, Trigaux D et al (2013a) SuFiQuaD: sustainability, ﬁnancial and quality evaluation of dwelling types. Belgian Science Policy (BELSPO), Brussels Allacker K, Debacker W, Delem L et al (2013b) Environmental proﬁle of building elements. OVAM, Mechelen Allacker K, Souza DM de, Sala S (2014) Land use impact assessment in the construction sector: an analysis of LCIA models and case study application. Int J Life Cycle Assess 19:1799–1809. https://doi.org/10.1007/s11367-014-0781-7 ASPEN (2015a) ASPENINDEX Regio België—Nieuwbouw (translated title: ASPENINDEX region Belgium—new construction). Antwerpen ASPEN (2015b) ASPENINDEX Regio België—Ombouw (translated title: ASPENINDEX region Belgium—renovation). Antwerpen Belgian Federal Government (2017a) Statistics Belgium—oil prices. http://statbel.fgov.be/nl/ statistieken/cijfers/energie/prijzen/gemid_8/. Accessed 12 Mar 2017 Belgian Federal Government (2017b) Statistics Belgium—building land prices. http://statbel.fgov. be/nl/statistieken/cijfers/economie/bouw_industrie/vastgoed/gemiddelde_prijs_bouwgronden/. Accessed 12 Mar 2017 Berghauser Pont M, Haupt P (2010) Spacematrix. NAi Publishers, Rotterdam 36 D. Trigaux et al.

CEN (2010) EN 15643-1 sustainability of construction works—sustainability assessment of buildings—part 1: general framework CEN (2011) EN 15978 sustainability assessment of construction works—assessment of environmental performance of buildings—calculation method CEN (2013) EN 15804:2012+A1 sustainability of construction works—environmental product declaration—core rules for the product category of construction products CEN (2015) EN 16627 sustainability of construction works—assessment of economic performance of buildings—calculation methods CREG (2015) Overzicht en evolutie van de elektriciteits- en aardgasprijzen voor residentiele klanten De Nocker L, Debacker W (2015) Annex: update monetisation of the MMG method (2014). OVAM, Mechelen Declercq K, Reumers S, Polders E, et al (2016) Onderzoek Verplaatsingsgedrag Vlaanderen 5.1 (2015–2016), Tabellenrapport (translated title: Research displacement behaviour in Flanders 5.1 (2015–2016), Tables report). Instituut voor Mobiliteit, Universiteit Hasselt, Diepenbeek Delhaye E, De Ceuster G, Maerivoet S (2010) Internalisering van externe kosten van transport in Vlaanderen (translated title: Internalisation of external cost of transport in Flanders). VMM, Mechelen Delhaye E, De Ceuster G, Vanhove F, Maerivoet S (2017) Internalisering van externe kosten van transport in Vlaanderen: actualisering 2016 (translated title: Internalisation of external cost of transport in Flanders: actualisation 2016). VMM, Aalst EC-JRC (2011) International reference life cycle data system (ILCD). In: Handbook— recommendations based on existing environmental impact assessment models and factors for life cycle assessment in a European context. Joint Research Centre (JRC) of European Commission—Institute for Environment and Sustainability (IES) European Environment Agency (2006) Urban sprawl in Europe, the ignored challenge. European Environment Agency, Copenhagen, Denmark Feist W, Schnieders J, Loga T, et al (2001) Energiebilanzen mit dem Passivhaus Projektierungs Paket (translated title: Energy balance with the passive house project package). Darmstadt Flemish Government (2011) Duurzame woningbouw—Vlaamse Maatstaf voor Duurzaam Wonen en Bouwen—versie 2.0 (translated title: Sustainable housing—Flemish tool for sustainable living and building—version 2.0). Departement Leefmilieu, Natuur en Energie (LNE) Flemish Government (2016) Regional urban planning regulation on rainwater management. https://www.ruimtelijkeordening.be/Verordeningen/Hemelwater. Accessed 20 Aug 2017 Flemish Government (2017) EPB requirements. www.energiesparen.be/epb/welkeeisen. Accessed 20 Aug 2017 Frischknecht R, Jungbluth N, Althaus H et al (2007) Overview and methodology—final report ecoinvent data v2.0, no 1. ecoinvent Centre, Dübendorf Ratti C, Baker N, Steemers K (2005) Energy consumption and urban texture. Energy Build 37:762–776. https://doi.org/10.1016/j.enbuild.2004.10.010 Salat S (2009) Energy loads, CO2 emissions and building stocks: morphologies, typologies, energy systems and behaviour. Build Res Inf 37:598–609. https://doi.org/10.1080/09613210903162126 Spon press (2015a) Spon’s external works and landschape price book 2015, 34th edn. AECOM, London Spon press (2015b) Spon’s civil engineering and highway works price book, 29th edn. AECOM, London Trigaux D (2017) Elaboration of a sustainability assessment method for neighbourhoods. PhD dissertation, KU Leuven Trigaux D, Allacker K, De Troyer F (2014a) Model for the environmental impact assessment of neighbourhoods. In: Passerini G, Brebia CA (eds) Environmental impact II. WIT Press, Ancona, Italy, pp 103–114 Trigaux D, Allacker K, De Troyer F (2014b) A simplified approach to integrate energy calculations in the life cycle assessment of neighbourhoods. In: Rawal R, Manu S, Khadpekar N (eds) Sustainable habitat for developing societies—book of abstracts. CEPT University Press, Ahmedabad, India, pp 55–55 2 Modelling the Influence of Urban Planning on the Financial … 37

Trigaux D, Oosterbosch B, Allacker K, De Troyer F (2015) A design tool to optimize solar gains and energy use in neighbourhoods. In: Cucinella M, Pentella G, Fagnani A, D’Ambrosio L (eds) Architectuur in (R)evolution—book of abstracts. Ass. Building Green Futures, Bologna, Italy, pp 305–305 Trigaux D, Allacker K, De Troyer F (2017a) Life cycle assessment of land use in neighborhoods. Procedia Environ Sci 38:595–602. https://doi.org/10.1016/j.proenv.2017.03.133 Trigaux D, Oosterbosch B, De Troyer F, Allacker K (2017b) A design tool to assess the heating energy demand and the associated financial and environmental impact in neighbourhoods. Energy Build 152:516–523. https://doi.org/10.1016/j.enbuild.2017.07.057 VMM (2015) Gemiddelde waterprijs 2015. https://www.vmm.be/data/gemiddelde-waterprijs. Accessed 24 Jan 2017 Trimble Inc. (2017) SketchUp. www.sketchup.com. Accessed 21 Mar 2017 VEA (2013) Energieprestatiecertificaten voor bestaande residentiële gebouwen in Vlaanderen— Formulestructuur (translated title: Energy performance certificate for existing residential buildings in Flanders—formula structure) VREG (2017) Electricity use of a household. www.vreg.be/nl/elektriciteitsverbruik-van-een-gezin. Accessed 28 Mar 2017 Chapter 3 Achieving Energy Efficiency in Urban Residential Buildings in Vietnam: High-tech or Low-tech?

Quang Minh Nguyen

Abstract Vietnam started to go for green rather late, officially in 2005 or 2006, shortly before Vietnam Green Building Council was established and the first legal documents paving the way for green building development to take root were drafted and then adopted. As the green building is a holistic concept and encompasses a wide range of specialisation, it seems that at the beginning, Vietnam chose energy —the most important component—to focus on before dealing with other measures in a comprehensive package of solutions. In terms of energy, most green, and energy-efficient buildings that Vietnam has constructed and been certified so far come from public and industrial building sector, such as schools, supermarkets, offices, showrooms and factories, while in housing which makes up the largest part of the country’s urban building market, this concept has not been properly developed, not only in quality but also in quantity. In order to provide more energy-efficient housing for the public and meet their very high demand, there are two options for architects—high-tech design and low-tech design—to propose in the local context. Furthermore, another possibility—combining the two tendencies —should be considered and discussed, because of its potential and flexibility in practice, as well as appropriateness in Vietnamese conditions.

Keywords Urban housing Á Green building Á Energy-efﬁcient architecture

3.1 Introduction

3.1.1 Urbanisation

As a developing country centrally located in Southeast Asia, Vietnam has been ranked among the fastest growing emerging economies in the world since 2000. International reports show that the urbanisation rate in Vietnam has increased steadily

Q. M. Nguyen (&) National University of Civil Engineering, 55 Giai Phong Road, Hanoi, Vietnam e-mail: [email protected]

© Springer Nature Singapore Pte Ltd. 2019 39 E. Motoasca et al. (eds.), Energy Sustainability in Built and Urban Environments, Energy, Environment, and Sustainability, https://doi.org/10.1007/978-981-13-3284-5_3 40 Q. M. Nguyen in recent years, estimated at 0.6–0.7% annually, from 27.89% in 2006 to 34.24% in 2016 (Statista 2017). The latest statistics reveal that 35.2% of Vietnam’s total population (33.99 million out of 96.51 million) live in cities and towns (World Bank 2018). This rate is forecast to reach 36.3% in 2020, 41.2% in 2030 and 48.6% in 2050 (Worldometer 2018). Today, there are 813 cities and towns across the country (Ministry of Construction 2018). Ho Chi Minh City (8.2 million residents) and Hanoi (7.4 million residents) are Vietnam’s two largest urban centres (World Atlas 2017), two special-class cities and also the two most important economic hubs in the South and in the North, respectively. The rapid urbanisation rate has caused many social problems and puts a burden on the current underdeveloped technical infrastructure systems. In this circumstance, housing has become a hot spot (Fig. 3.1).

3.1.2 Urban Housing

Urban housing development in Vietnam was ignited in the early 1990s with a number of legal documents and supporting policies, such as Housing Ordinance in 1991, Land Law in 1993, Building Ownership and Certificate of Individual Land Use Rights in 1994, later reinforced with National Orientation for Housing Development in 2004, Housing Law in 2005 and Real Estate Law in 2006. Today, after almost three decades, housing is reported to make up no less than 70% of all the built floor area in Vietnamese cities, and the actual demand for new housing is estimated at over 500,000 units per year after 2020 or 12.5 million m2 should be annually provided, compared to 325,000 units (6.5 million m2) per year in 2010–2020 and 275,000 units (4.1 million m2) per year before 2010, as the housing area standards of 25, 20 and 15 m2 per person should apply respectively (UN Habitat Vietnam 2014; Ministry of Construction 2011). This demand has often exceeded the supply capacity of the housing construction and building industry in Vietnam over the last 10 years, which was severely affected by the world’s economic and financial crisis in 2008 and again by the national frozen real estate market from 2010 until 2014/2015. There is a mixture of three to six housing types in almost all of the current living quarters in major cities like Hanoi and Ho Chi Minh City. This trend began in the late 1980s, after the centrally planned economy was abolished and the long-awaited economic reform also marked a new beginning for urban housing, as it would no longer be a sector monopolised by state-owned building corporations as previously (1954–1986) and the individual property ownership could eventually be recognised for the first time by the Government in 1994. Since 2010, this kind of mixing has accelerated. In Hanoi, for example, there are 12 housing types listed in chrono- logical order, as the city expanded: traditional shophouse, traditional garden house, French colonial villa, French colonial shophouse, Soviet collective block of flats, self-built family house (into a row), self-built villa, social rental house, mini-apartment building, standard villa, standard row house and standard apartment building (multistorey or high-rise). Slum is excluded, as it has never been officially regarded as formal housing and the city authority is trying to replace it with new 3 Achieving Energy Efficiency in Urban Residential Buildings … 41

Fig. 3.1 Vietnam administrative map (mainland) and its city-and-town system. The two archipelagos of Paracel (Hoang Sa) and Spratly (Truong Sa), which Vietnam claims national sovereignty, are not shown on the map due to the frame dimensions. (Nation Online Maps 2018) 42 Q. M. Nguyen social housing by 2020. Among these 12 housing types, self-built family house was reported in 2014 by UN Habitat Vietnam to make up the largest part (estimated at 70%). Building standard apartments (for both high-end and low-end users) is the fastest-growing sector in mega-cities, because it has been determined a major housing pattern in the future and constructed on a large scale to meet the urgent needs for housing of millions of both permanent dwellers and new immigrants rushing from the countryside as well as from small towns. A very similar situation is noted in Ho Chi Minh City. Much of urban housing in Vietnam’s two mega-cities fails to meet the new housing quality and floor area per capita standard proposed by the Ministry of Construction, particularly in central districts, where the building density is particularly high. Lack of daylight and natural ventilation, extremely low green area per capita, inappropriate use of building materials, abuse of air condi- tioners, negative influences of polluted air and noise from traffic congestions, insufficient water supply and inadequate sanitation are common housing problems in these areas (Fig. 3.2).

3.1.3 Green Building

Energy efficiency has been, and continues to be, the most important component in a truly green building, because both energy production and energy consumption are largely involved in and responsible for the global warming, environmental crisis and other social problems arising from human daily activities. In the face of fossil fuel exhaustion, environmental pollution and increasing demand for energy as a consequence of population boom, rapid urbanisation and machine-based modern life, energy efficiency has been regarded as a must, not only in new building projects but also for the conservation of historic buildings worldwide. However, design for energy efficiency seems to take place mostly in developed countries in North America, Western Europe, Japan, Australia, New Zealand and Singapore. In other parts of the world, including Vietnam, this trend has not yet (or just) taken its root and people have therefore a long way to go to keep pace with the world’s progress and achievement. Established in 2007 and officially recognised by the Ministry of Construction in 2009, Vietnam Green Building Council (VGBC) is responsible for the promotion and development of green buildings in Vietnam by setting up a comprehensive green building rating system and encouraging green building design as well as awarding green building projects across the country. Lotus is a green building certificate granted by VGBC. The 2018 database of VGBC consists of 43 green building projects, divided into eight categories and five ranks as follows: Category 1 (Industrial buildings): 7 Category 2 (Education buildings): 8 Category 3 (Tourism buildings): 1 Category 4 (Office buildings): 8 3 Achieving Energy Efficiency in Urban Residential Buildings … 43

Fig. 3.2 a Housing development in central Ho Chi Minh City (Photos: taken in 2018 by Quang Minh Nguyen). b Housing development in central Hanoi (Photos: taken in 2018 by Quang Minh Nguyen)

Category 5 (Service buildings): 1 Category 6 (Commercial buildings): 5 Category 7 (Cultural buildings): 1 Category 8 (Residential buildings): 12 Total: 43 44 Q. M. Nguyen

Rank 1 (Platinum certificate): 3 Rank 2 (Gold certificate): 0 Rank 3 (Silver certificate): 9 Rank 4 (Certificate): 5 Rank 5 (Provisional certificate): 26 Total: 43 (Vietnam Green Building Council 2018) It is noticeable that residential buildings (category 8) have not yet been largely constructed for green building certificates: out of 12 residential buildings listed, only one has been officially certified while the other 11 are still under examination or partially evaluated—thus provisionally certified. Similarly, most of the buildings (31 out of 43—over 72%) acquire lower ranks. For the next 10 years, it will be necessary to have a greater number of green buildings, particularly in housing, and higher standards (platinum and gold) should be notched up. The statistics from VGBC also reveal that by the end of 2017, 116 buildings and building projects in Vietnam had been awarded green building certificates by one of the following four bodies: VGBC with Lotus, USGBC with LEED, Building and Construction Authority (BCA) Singapore with Green Mark and International Financial Corporation (IFC)/World Bank with EDGE (Excellence in Design with Greater Efficiency). This growth rate is quite positive for the first ten years of developing green buildings in Vietnam (2007—2017), but still incomparable to 2,100 green buildings that Singapore had achieved by 2014. The more ambitious plan for BCA to strive for is to have 80% of all buildings qualified with Green Mark by 2030 (BCA, 2014). Vietnam must make greater efforts in the coming years, if the country does not want to lag behind in the ASEAN in terms of developing green buildings (Fig. 3.3). Certainly, the VGBC database is not a complete list, because there are other green buildings in reality which have not yet been registered for official recognition. Some buildings have become quite well known, even won national/international awards,

Fig. 3.3 Green building growth in Vietnam from 2007 to 2017 (accumulative ﬁgures) (Data: VGBC, 2018—Photos: taken in 2018 by Quang Minh Nguyen) 3 Achieving Energy Efﬁciency in Urban Residential Buildings … 45

Fig. 3.4 Suoi Re Community House (left) and S-House 2 as two of the first green buildings in Vietnam. Internationally recognised with grand awards but not yet nationally certified. (Photo: taken in 2012 by Quang Minh Nguyen) and (Vo Trong Nghia Architects 2018) such as Suoi Re Community House in Hoa Binh province designed by Architect Hoang Thuc Hao (1 + 1 > 2 Office) or Palm tree House (S-House 2) in Long An Province developed by Architect Vo Trong Nghia (VTN Architects) (Fig. 3.4).

3.1.4 Energy Consumption

A high level of energy consumption (and CO2 emission as well) can be seen in many residential buildings in urban agglomerations like Ho Chi Minh City and Hanoi. Two individual investigations into household energy consumption were conducted in 2016 with 110 case studies for row houses and 35 case studies for apartments, because these are the two most typical housing patterns in central Hanoi in terms of thermal discomfort and excessive energy use as a consequence of inappropriate design and construction. The results show that household energy consumption immensely depends on the following three main factors: orientation, design concept and purpose of use. A small number of case studies with the same (or similar) input data, such as gross ﬂoor area, family size, location, orientation, etc. have been selected and analysed for comparison purposes. Orientation: A row house and an apartment facing Southeast consume less energy than the same size house and apartment facing Southwest in the same area, 47.7% and 26.8%, respectively. South and Southeast are two optimal directions for residential buildings in Hanoi and the Red River Delta (North of Vietnam), because residents can make full use of warm morning sunlight and cool wind in summer months while avoiding harmful solar radiation in the afternoon and cold winter wind from late November to mid-March. Bio-climatically, a house or an apartment facing West or Southwest are rather uncomfortable. But choosing a better orientation is sometimes not possible in planning. In this case, additional measures must be taken, either architecturally or technically (or both, if necessary) to help min- imise the negative inﬂuences of the weather conditions on the indoor environment. 46 Q. M. Nguyen

Design concept: Between two row houses or two apartments of the same size facing Southwest as one of the two most disadvantageous directions, the one with simple but efﬁcient design solutions against heat and solar radiation consumes 25.5% less energy (row house) or 22.0% less energy (apartment) than the other that goes without applying such solutions as suggested by architects. Business and service activity: Among three row houses facing the same direction (Southwest), the one for living only consumes 15.4% less energy than the one with a small shop and simple service and 35.3% less energy than the one with a large shop and special service (Author, 2016) (Table 3.1). Generally, there is no rule of the ups and downs in the household energy consumption, because the family’s activities change from time to time or some events may happen suddenly, such as going on holiday or business, welcoming visitors and guests, extreme weather conditions, etc. But it is apparent to notice that the energy consumption in this case study increased over the years. The energy demand becomes much higher in summer months (May to August) due to very hot weather and longer time of air conditioning. Similarly in January or February, when the New Year’s Holiday comes, cooking and washing can be energy-consuming. In February and March, when the air humidity amounts to 95– 100% because of spring drizzles, drying is a major demand, as dehumidiﬁers are switched on 24 h a day. Another survey conducted on a much larger scale by the National Institute of Energy in 2016 reveals that the energy consumption per capita rose continually from 2010 to 2015 (European Chamber of Commerce in Vietnam 2017). This

Table 3.1 Comparative study on household energy consumption in Hanoi (Survey: undertaken in 2016 by Quang Minh Nguyen) Factors and cases Energy consumption (kWh/m2a) Row house Apartment A. Orientation (with proper design, no shop) A1. Southeast 23 41 A2. Southwest 44 56 Assessment (A1 compared to A2) 47.7% saving 26.8% saving B. Design concept (same direction—Southwest) B1. With solutions against heat and solar radiation 35 59 B2. Without solutions against heat and solar radiation 47 73 Assessment (B1 compared to B2) 25.5% saving 19.2% saving C. Business and service (same direction—Southwest) C1. No activity (with no shop) 44 – C2. With activity (small shop, simple service) 52 – C3. With activity (large shop, specific service) 68 – Assessment (C1 compared to C2) 15.4% saving Assessment (C1 compared to C3) 35.3% saving 3 Achieving Energy Efficiency in Urban Residential Buildings … 47 upward trend is in accordance with the individual study results shown in Table 3.2. The values in Fig. 3.5 are calculated on an average basis, between urban and rural areas, thus lower than those in Table 3.2 from Hanoi as a mega-city, considered for 2013, 2014 and 2015 as 3 years in common between the two surveys. The main problem to be tackled for housing development in major cities in Vietnam, especially in Hanoi and Ho Chi Minh City, is energy efficiency that can be achieved by applying a number of planning strategies, design concepts and technical solutions. Planning strategies deal with optimal orientation which may be difficult to select for various reasons. For example, one of the two rows of houses in a street running from the North to the South has to face West. Built on a limited land plot, that row house cannot be rotated to face south like a villa in the middle of a large garden. In this case, it is highly recommended to focus on housing design concepts by reorganising the conventional room layout into a new one in view of energy (form follows energy apart from function) and by restructuring the building envelope, so that the external walls and the roofs can be better protected from solar radiation and excessive heat. For the next step, with several technical solutions that will help optimise cooling as well as heating and reduce the dependence on fossil fuel energy sources (coal, oil and gas), it is possible to enhance energy efficiency

Table 3.2 Household electricity consumption in one typical middle-class family in Hanoi (parents and two children) (Data: collected and analysed in 2018 by Quang Minh Nguyen) Month Annual energy consumption (kWh) 2013 2014 2015 2016 2017 January 554 663 678 833 959 February 593 628 645 625 983 March 558 569 580 638 1,103 April 539 576 593 693 770 May 602 611 587 591 630 June 775 795 844 813 689 July 646 861 787 945 880 August 535 588 631 907 893 September 620 696 915 878 679 October 603 642 705 738 666 November 512 525 531 553 603 December 532 558 527 583 589 Total (kWh) 7,069 7,712 8,023 8,797 9,444 Gross ﬂoor area (m2) 160 160 160 160 160 Energy consumption (kWh/m2a) 44 48 50 55 59 Number of family members 4 4 4 4 4 Consumption per capita (kWh/ 1,767 1,928 2,006 2,199 2,361 person) Quantiﬁcation x 1.09x 1.13x 1.24x 1.33x 48 Q. M. Nguyen

Fig. 3.5 Household energy consumption per capita in Vietnam from 2010 to 2015 (Data: European Chamber of Commerce in Vietnam 2017—Graphic: drawn by Quang Minh Nguyen, 2018

considerably. In other words, CO2 emission into the atmosphere can be controlled and global warming will not be going on too fast. To solve this problem, architects have several options in two directions: high-tech solutions and low-tech solution, which will be discussed later at the end of this book chapter. The objectives of developing more energy-efficient housing concepts should be set and highlighted as follows: • To improve indoor thermal comfort for building occupants; • To use energy more efficiently and to promote energy efficiency in modern housing and building design: the same energy consumption for greater thermal comfort or the same thermal comfort with less energy consumption; • To enhance the awareness of energy efficiency and sustainability among the public and decision/policy makers as well as project managers; • To revitalise some of the best traditional housing concepts towards a more comfortable built environment; • To pave the way for a new revolution in urban housing development.

3.2 Energy and Energy Efﬁciency in Vietnam

3.2.1 Strengths

Entirely located in the tropical climate zone and with a long coastal lines (over 3,200 km), Vietnam enjoys a huge potential and has very good opportunities to exploit renewable energy sources, especially solar energy and wind energy which are expected to replace conventional forms of energy step by step until 2030. Solar 3 Achieving Energy Efficiency in Urban Residential Buildings … 49 energy can be produced with high solar energy intensity and annual sunshine hours, as given in Table 3.3 and in Fig. 3.6. Apart from solar energy development plans and construction of two photovoltaic panel factories in Vietnam to reduce the equipment fabrication price and encourage people to start using clean energy by installing solar energy systems at homes, the Government and the Ministry of Industry and Trade have cooperated with Denmark and Germany as two world leaders in wind power development to build some wind power stations along the coastal line. The first two large-scale projects in Ninh Thuan province and Bac Lieu province are underway. In addition, tidal energy can be exploited but the potential has not yet been explored and/or assessed. In green building and clean energy development, strategies and policies play a vital role, particularly for a country that has just started to go for green like Vietnam. Understanding this, the Government and the Ministry of Construction have prepared some laws to pave the way for energy efficiency to become an obligatory requirement in both industrial and civil construction, as well as for regenerative energy (primarily solar energy and wind energy) to be more popular among city inhabitants. The most significant legal documents include Energy Saving and Energy Efficiency Law (No. 50/2010/QH12) adopted in 2010 by the National Assembly and the National Energy Efficiency Building Code (QCVN09/ 2017/BXD) proposed in 2016 by the Ministry of Construction and approved in 2017 by the Prime Minister. Energy Saving and Energy Efficiency Law is applicable to all households, individuals and organisations that use energy for any purpose in Vietnam, consisting of 12 chapters with 48 articles. Key articles include: • Article 15: Recommendations for energy saving and energy efficiency in housing construction and operation; • Article 27: Regulations for energy efficiency in everyday household activities; • Article 37: Use of standard energy-efficient equipment and devices in line with energy labelling; • Article 38: Establishment and announcement of energy efficiency standards; • Article 42: Optimisation of design and utilisation of renewable energy sources, development of innovative energy technologies.

Table 3.3 Potential of solar energy production in Vietnam (European Chamber of Commerce in Vietnam Ofﬁce 2017) Region Sunshine Solar radiation Assessment hours per intensity of potential year (h) (kWh/m2.day) North (with Hanoi, Hai Phong, Nam 1,500–1,700 3.8–4.7 Fairly high Dinh, Ha Long) Central (with Da Nang, Hue, Quy 1,800–2,100 4.5–5.3 High Nhon, Nha Trang) South (with Ho Chi Minh City, Bien Hoa, 2,000–2,600 5.1–5.6 Very high Vung Tau, Can Tho) 50 Q. M. Nguyen

Fig. 3.6 Map of solar radiation intensity and its potential in Vietnam (Vu Phong Solar Power Company 2017) 3 Achieving Energy Efﬁciency in Urban Residential Buildings … 51

Meanwhile, the National Energy Efficiency Building Code (upgraded in 2017 on the basis of the previous version in 2013) is applicable to all buildings with gross floor area of 2,500 m2 or above, no matter what kind of building they are. Important requirements and technical specifications include: • Building envelope (roof, external wall, windows and doors); • Overall Thermal Transferred Value (OTTV) for different building elements; • Solar Heat Gain Coefficient (SHGC) for building elements; • Window/Wall Ratio in different orientations; • A factor for different sun-shading elements in different building orientations; • Natural and mechanical ventilation; • Cooling coefficient of performance; • Technical specifications for cooling towers and condensers; • Insulation for cooling systems and thickness of insulation layers; • Daylighting and minimum illuminance; • Productivity and performance of lighting devices; • Monitoring and automation of street lighting and lighting for public places; • Energy use and electricity distribution systems; • Measurement of energy consumption and adjustment of power; • Warm water systems; • Insulation for warm water systems and thickness of insulation layers; • Maximum U value for external walls: 1.8 W/m2K; • Maximum U value for roofs: 1.0 W/m2K; • Maximum OTTV value for external walls: 60 W/m2; • Maximum OTTV value for roofs: 25 W/m2 (Ministry of Construction 2017). These are two effective instruments for practising energy efficiency in buildings. However, the floor area of 2,500 m2 normally applies to medium-size public buildings and large-scale high-rise residential buildings (20 or more apartments per floor). Single-family houses have not yet taken into account, although they exist in a huge number and consume so much energy. Vietnam Green Building Council (VGBC) has developed several detailed rating systems for green buildings and certificates in general, and in housing in particular, with two categories: multifamily and single family. These rating systems are accompanied by intelligible technical manuals (Table 3.4). The importance of energy efficiency is demonstrated in the highest scores (31 out of 118 points for multifamily buildings and 29 out of 100 points for single-family buildings). In both cases, energy efficiency has been broken down into eight sub-criteria with corresponding scores. This rating system is quite clear to understand and easy to use for a preliminary qualitative self-assessment of building performance at the design stage. Since 2014, Vietnam Association of Architects (VAA) has used a set of criteria to evaluate and grant awards the best designs within the framework of the annual National Green Building Competition (Table 3.5). 52 Q. M. Nguyen

Table 3.4 a VGBC rating system (and scorecard) for multifamily residential buildings in category Energy (Vietnam Green Building Council 2012). b VGBC rating system (and scorecard) for single-family residential buildings in category Energy (Vietnam Green Building Council 2012) a No. Criteria and sub-criteria Maximum point Given point 1 Energy 31 1.1 Passive design Prerequisite 1.2 Total building energy use 14 1.3 Building envelope 4 1.4 Natural ventilation and air conditioning 6 1.5 Artificial lighting 3 1.6 Energy monitoring 1 1.7 Lifts 1 1.8 Renewable energy 2 2 Water 13 3 Materials 9 4 Ecology 9 5 Waste and pollution 7 6 Health and comfort 13 7 Adaptation and mitigation 10 8 Community 6 9 Management 12 10 Innovation 8 Total 118 b 1 Energy 29 1.1 Passive design 5 1.2 Building envelope 4 1.3 Home cooling 6 1.4 Artificial lighting 3 1.5 Water heating 2 1.6 Energy-efficient appliances 3 1.7 Energy monitoring 1 1.8 Best practice credits 5 2 Water 12 3 Materials 14 4 Local environment 17 5 Heat and comfort 14 6 Community management 10 7 Innovation 4 Total 100 3 Achieving Energy Efficiency in Urban Residential Buildings … 53

Table 3.5 Green architecture criteria with focus on energy set up by vietnam association of architects (Vietnam Association of Architects 2017) No. Criteria and sub-criteria Maximum Given point point 1 Sustainable site 15 2Efficient use of natural resources and energy 40 2.1 Complying with National Technical Standards and Prerequisite current legal documents regarding land-use indicators, condition energy use, water use and material use in construction 2.2 Efficient exploitation and use of land in 12 construction 2.3 Efficient exploitation and use of water 12 resources 2.4 Efficient exploitation and use of energy 10 2.4.1 Finding planning solutions, proposing 7 architectural designs, making good selections of building materials, applying technology and installing equipment to ensure: 2.4.1.1 Efficient use of natural energy sources 4 (sun, geothermal heat, etc.), exploitation and making full use of other renewable energy sources which are locally available and easy to regenerate, etc. Using renewable energy at least 5% of the total energy consumption within a building 2.4.1.2 Making full use of daylight and natural 3 ventilation 2.4.2 Using monitoring and controlling systems in 3 order to manage energy use in a building towards minimisation of energy consumption (such as EMS—Energy Management System, etc.) 2.5 Exploitation and efficient use of materials 6 (building envelopes and interior design) 3 Indoor air quality and environmental quality 13 in urban or rural areas 4 Advanced architecture and identity 17 5 Social sustainability and humanity 15 Total 100

Compared to the rating systems proposed by VGBC, VAA’s criteria are more complex and comprehensive, because social sustainability has also been integrated. Energy is directly given 10 out of 100 points only, but the actual score for energy amounts to 31, with 21 additional points from several sub-criteria related to energy in Categories 1 and 3. 54 Q. M. Nguyen

3.2.2 Opportunities

In the era of globalisation, it is possible to make full use of international cooperation, including the fields of building science and energy technology, as an outstanding strength for the future development of energy efficiency in Vietnam. The country has good opportunities to develop renewable energy based on its high potential and energy-efficient buildings with innovative energy technologies directly transferred from the EU and/or the USA. Germany and Denmark have helped Vietnam build wind turbines in some coastal provinces in the south while Winrock International Corporation from the USAID has been a reliable partner in bringing solar energy to a larger number of people in Vietnam, especially those living in the countryside, and in launching more action programmes for the communities easily affected by climate change so that they can mitigate the impacts of unsustainable development on the environment. In the future, such cooperation activities should be consolidated and new international organisations will be invited to give a further boost to energy-efficient housing in Vietnam.

3.2.3 Weaknesses

High-tech energy production remains a weakness in Vietnam, but that problem can be solved with cooperation and technology transfer as aforementioned. Incomplete understanding of green building design among professionals and lack of awareness of sustainable development among the community are the real hurdles for Vietnam to overcome on the way to sustainability and energy efficiency in urban housing. In fact, an energy-efficient building can only be achieved by applying a wide range of design concepts and a combination of technical solutions. However, many architects in Vietnam, even those with more than 10 years of work experience, still imagine a green home simply as a row house with a green roof, an apartment with a green façade, or a villa in the middle of a large land plot with so many shade trees planted all around the house. They consider an energy-efficient building a building with photovoltaic panels (for electricity) and a solar collector (for warm water) installed on the roof. As a consequence, inappropriate design solutions can be found in numerous housing concepts, for example, air-tight corridors and bedrooms without windows (Fig. 3.7). In Artemis Tower—one of the best apartment buildings in Hanoi designed by PTW as a leading Australian company—14 out of 20 apartments in one typical floor plan have one bedroom and one kitchen with inadequate daylight and natural ventilation due to their location, just connected to the open air through a long and narrow loggia (1 m Â 4.2 m), as marked with red box (bedroom) and yellow box 3 Achieving Energy Efficiency in Urban Residential Buildings … 55

Fig. 3.7 Analyse of design shortcomings in Artemis tower as one of the best deluxe apartment buildings in Hanoi (Background ﬂoor plan: Artemis Project Management Unit, 2017—Design: analysed by Quang Minh Nguyen, 2018)

(kitchen) in Fig. 3.7. One of the three staircases has no daylight and air exchange which is highlighted with a purple box. A majority of people think about green building/housing almost in the same way as architects do. They admit that green/energy-efﬁcient architecture sounds very nice, but they ﬁnd the price so high that they can hardly (or never) afford it. That is why they feel reluctant to purchase a truly greenhouse or an apartment designed in compliance with the latest energy-saving standard. Instead, most of them choose other options of lower building quality or energy standard that fall within their payability. In their opinion, air conditioning is the only way to ensure thermal comfort in the city with a very high building density and huge impacts of urban heat island like Hanoi or Ho Chi Minh City. People tend to abuse air conditioning, even when it is not necessary to run a mechanical ventilation system. While specialists recommend that the room temperature should be set at 25 or 26 °C for the outside air temperature at 34 or 35 °C at night in the summer months, many occupants select the room temperature mode at 20 °C or even lower from 10 pm to 6 am the following day (8 hours long) which of course consumes a huge amount of electric power. It is important to know that only by turning off all light bulbs that might be unnecessary for use within one hour did the Earth Hour campaigns help Vietnam save 471,000 kWh in 2017 and 485,000 kWh in 2018 (Vietnam News 2018).

3.2.4 Threats

Today, in electricity production, Vietnam still has to depend so much on fossil fuels, such as coal, oil and gas which cause serious environmental pollution. Heat power plants, for this reason, have not been recommended. Hydropower is not as abundant as in the 1970s or 1980s. On-going climate change has resulted in a much lower of 56 Q. M. Nguyen

Fig. 3.8 Energy production in 2015 and energy development scenarios for the future in Vietnam (Data: Vietnam Institute of Energy, 2016—Graphic: drawn by Quang Minh Nguyen, 2018) rainfall and downstream flow. In addition, China goes on to build huge dams in the upper tributary of the Red River (in Yunnan province) as the most important source of water and hydropower to the northern part of Vietnam. The recent energy production structure (as of 2016) and the future energy development plan (for 2020 and 2030) show that clean energy has not yet been properly developed in Vietnam (Fig. 3.8), even when Vietnamese energy experts and some international organisations advised the Government and the Ministry of Industry and Trade to invest in clean energy instead of low-tech power production imported from China. In 2014, a decision to build and put two nuclear power plants (4,600 MW) into use by 2030 with Russian nuclear power technology in Ninh Thuan province (Vietnam Electricity Corporation 2016)—not so far from the wind power plant supported by Denmark—raised a great public concern, because of a high risk to the local eco-systems and living environment, just 3 years after the nuclear catastrophe in Fukushima (Japan) happened. Although the Government cancelled that contro- versial plan in 2016 as a result of the public strong reaction, Vietnam will have to face energy insecurity and environmental challenges as far as no bold step towards renewable energy has ever been made. As indicated in Fig. 3.8, three renewable energy sources (solar energy, wind energy and biomass) will contribute only 3.3% to the total energy production in 2020 and increase to 16.1% in 2030—a considerable improvement. However, in view of the urgent need for more renewable energy, this growth rate is still insufficient. Thermo-electricity from fossil fuels (coal, gas and oil) will make up a large part of the energy market in 2020 and even until 2030: approximately 60% (Vietnam Institute of Energy 2016). The treatment of pollution caused by the production of thermoelectricity will cost a great deal of money and put an immense financial burden on a developing country like Vietnam. After all, the threats can be seen in the intention of developing nuclear power technology and the policy of using fossil fuels in the electricity production. 3 Achieving Energy Efficiency in Urban Residential Buildings … 57

3.3 Energy Efﬁciency in Urban Housing in Vietnam—The First Two Examples

Building accounts for 36% of the nation’s total energy consumption and 33% of the CO2 emission a year (Vietnam News 2017). Therefore, achieving energy efficiency in building should be regarded as a key point in reducing both energy consumption and carbon emission in Vietnam and thereby securing sustainable urban development. In the building sector, housing should be chosen to apply energy efficiency first, because housing is the largest building market in the city—70% of the total built floor area (UN Habitat Vietnam 2014) with 689 million m2 of the entire urban housing stock as of 2014 (World Bank 2015) and people spend normally at least half a day staying in their houses or apartments. Mulberry Lane is one of the first housing projects in collaboration with CapitaLand Singapore and became the first residential project in Hanoi to receive a BCA Green Mark award in 2014 for excellence in housing design (overall band score of 79 out of 140 points) along with two national awards that same year for green building performance. All the five dwelling blocks A, B, C, D and E have been designed in view of tropical architecture with three different housing patterns, of which three blocks B, C and D in the middle have the same floor plan concept (Figs. 3.9 and 3.10). Four essential factors—daylight, cross ventilation, greening and stunning view—are maximised. Over 80% of the 1,487 apartments are regarded as ‘optimal’ in terms of daylight and cross ventilation, hereby reducing energy consumption for artificial lighting and air conditioning. RSP Architects 2015`` to the end of the paragraph and put into parentheses (RSP Architects 2015). In addition, the following solutions are applied to enhance energy efficiency: • Cooling with vegetation and vapour; • Choosing an optimal orientation (facing North and South); • Calculating an appropriate distance between two buildings (at least 20 m); • Designing a layer of insulation for roof and all external walls, especially in the West direction; • Double-glazing for all windows and doors open to balconies; • Using solar energy for warm water systems; • Installing energy-efficient HVAC systems and equipment, including lighting; • Raising public awareness of saving energy as daily behaviour. As a result of such a combination, the project can save up to 3.6 million kWh of electricity a year (CapitaLand Singapore 2015). The typical floor plan, in block B for instance, demonstrates a careful consideration in design for thermal comfort based on daylight and cross ventilation, applicable to all main apartment rooms (living room, bedrooms, dining room and kitchen). Instead of having 15–20 apartments per floor as often seen in other high-rise apartment building projects, a Mulberry Lane dwelling unit accommo- dates 10 or 13 apartments per floor only. Therefore, daylight and airflow can 58 Q. M. Nguyen

Case study 1 - Mulberry Lane project (RSP Architects, 2015) Project information: Site: Mo Lao new town - Ha Dong district, Hanoi City Site area: 2.4 ha Population: approximately 6,000 Hanoi Number of apartment buildings: 5 blocks (from 27 to 35 storeys) Gross floor area: 235,850 m2 Number of apartments: 1,487 Types of apartments: 45 m2, 80 m2, 88 - 114 m2, 123 - 127 m2 and 130 - 154 m2 Date of completion: 2014 Project developers: Hoang Thanh Building Corporation (Vietnam) and CapitaLand (Singapore) Building consultant: RSP Archi- tects Co. Ltd (Vietnam) Certification of VGBC green building: No Recognition of green building design: Yes

Case study 2 - Eden villa (XYZ Design Saigon, 2017) Project information: Site: Thao Dien ward, District 2, Ho Chi Minh City Ho Chi Minh City Site area: 450 m2 Type: Villa - single family house Building height: Three storeys Date of completion: 2017 Building consultant: XYZ Design Saigon

Fig. 3.9 Project sites for energy-efficient housing (Background map: Wikipedia 2018) penetrate into some rooms deep inside the buildings, as well as into some parts of the T-shaped or H-shaped corridors. The second example of energy-efficient housing comes from Ho Chi Minh City: Eden Villa. This is a new villa, designed exactly in tropical style, with a simple but elegant philosophy of maximising privacy, inside out view and most notably indoor comfort based on daylight and cross ventilation which have been successfully applied to the previous example and even more clearly reflected in this case. 3 Achieving Energy Efficiency in Urban Residential Buildings … 59

Fig. 3.10 a Aerial view of Mulberry Housing Project (Ha Dong District—Hanoi City) (RSP Architects Vietnam Co. Ltd. 2015). b Master plan of Mulberry Housing Project (Ha Dong District —Hanoi City) (RSP Architects Vietnam Co. Ltd. 2015). c Typical ﬂoor plan of Block B—Mulberry Housing Project (Ha Dong District—Hanoi City) (RSP Architects Vietnam Co. Ltd. 2015) 60 Q. M. Nguyen

Fig. 3.11 a Front view of Eden Villa (District 2—Ho Chi Minh City) (Archdaily Journal of Architecture 2017). b Back view of Eden Villa (District 2—Ho Chi Minh City) (Archdaily Journal of Architecture 2017). c Ground floor of Eden Villa (District 2—Ho Chi Minh City) (Archdaily Journal of Architecture 2017). d First floor of Eden Villa (District 2—Ho Chi Minh City) (Archdaily Journal of Architecture 2017). e Solar protection structure of Eden Villa (District 2— Ho Chi Minh City) (Archdaily Journal of Architecture 2017). f Cross section of Eden Villa (District 2—Ho Chi Minh City) (Archdaily Journal of Architecture 2017) 3 Achieving Energy Efficiency in Urban Residential Buildings … 61

Fig. 3.11 (continued)

Meanwhile, outdoor comfort can be secured with cooling effect from the swimming pool and shading from the overhang frame covered with creeping plants. Large blind windows protect the upper floor rooms from solar radiation. The building has recently been nominated for a grand award—Archdaily’s Building of the Year 2018 (Figs. 3.9 and 3.11). The level of energy efficiency will soon be verified because just after a few months in operation, the villa has not yet been energy-audited. 62 Q. M. Nguyen

The following solutions are applied to enhance energy efﬁciency: • Cooling with vegetation and evaporation (from the swimming pool at the back); • Maximising daylight and cross ventilation with four elevations open to the private garden, open structure; • Using sun-shading elements on the west façade; • Installing energy-saving lamps and electrical devices.

3.4 Discussion

3.4.1 An Open Question: High-tech or Low-tech to Achieve Energy Efﬁciency?

Today, architects in Vietnam will sooner or later have to answer the following two questions: Question 1: Is green building (or energy-efﬁcient housing) just accessible (or available) to high-income people? and Question 2: Do low-income and marginalised/underprivileged groups have any opportunities to own green/energy-efﬁcient houses/apartments?

Professionals will soon have to choose between high-tech pathway and low-tech pathway. Both concepts must be developed first and foremost on spatial and structural optimisation as a prerequisite condition in design, because no energy efficiency will ever be achieved without a good design concept, both architecturally and structurally considered. Then, technical solutions will only play a supporting role. High-tech housing concept means a house or an apartment that must fulfil the requirements for sustainability and energy efficiency with innovative (and usually expensive) techniques and equipment, such as triple glazed windows, smart double skin, automatically adjustable sun-shading systems or high-performance photovoltaic panels which may increase the housing price to 1,500 or 2,000 USD/m2. These cutting-edge technologies can only be found in many office buildings, such as Deutsches Haus (German House) designed by GMP Architekten in Ho Chi Minh City or VietinBank Tower designed by Norman Foster and Associates in Hanoi. In housing, such high-tech solutions have not been applied so far, mainly due to the very high total building cost—so high that very few house owners can afford. Another reason is that most Vietnamese architects are not so competent in high-tech design. Those who have a good and complete understanding of environmental engineering in architecture often graduated with international qualifications and go abroad to work. Low-tech housing concept, on the contrary, refers to an investigation into planning and building experience of the past and then an adaptation of such traditional experience to modern housing construction will be implemented. This second option proves to be practical, since the solutions are simple but efficient, and most 3 Achieving Energy Efficiency in Urban Residential Buildings … 63 importantly not expensive. This is absolutely appropriate for low-income or other underprivileged groups, exactly what Architect Anna Heringer has done for Indian and Bangladeshi people or Architect Diébédo Francis Kéré has helped villagers realise a better community life in the most remote areas in his home country Burkina Faso. Both architects used ecological and locally available building materials, for instance bamboo, wood, earth, stone, straw, etc. with manual building techniques and mobilised local workforce for construction. As a result, the construction did not cost the local communities so much money and they can enjoy nice and new libraries and schools designed in view of environment-friendly and energy-efficient building. In Vietnam, Architect Hoang Thuc Hao has been recognised in recent years as a pioneer in architectural design for the community and honoured with several awards for his contributions to the thriving social development. However, most of his works in this trend are public buildings, not housing. Learning from the past house planning and building experience is an essential part of applying low-tech design to modern building in order to ensure energy efficiency as well as sustainability. There are three housing patterns to explore:

Fig. 3.12 a Cross section of a typical traditional shophouse (Old Quarter, Hanoi, Vietnam) with daylight and cross ventilation effects (Graphic: drawn by Quang Minh Nguyen, 2018). b Cross section of a typical colonial shophouse (French Quarter, Hanoi, Vietnam) with daylight and cross ventilation effects (Graphic: drawn by Quang Minh Nguyen, 2018) 64 Q. M. Nguyen

Table 3.6 Low-tech design for thermal comfort and energy efficiency learnt from the past housing concepts (Data: analysed by Quang Minh Nguyen, 2018) Solution Traditional Colonial shophouse Colonial villa shophouse Daylight gain Through the front Through the front All four sides (single villa) side and three to side and two to three or three sides (duplex villa) four courtyards courtyards Maximisation of Through courtyards Through courtyards Through opposite large cross ventilation and opposite and opposite narrow windows narrow windows windows Thermal Double layer of ∙ Double layer of ∙ Double layer of insulation for traditional factory-made factory-made roofing tiles roofs and external hand-made roofing roofing tiles with air space underneath walls tiles ∙ Gypsum-straw ∙ Gypsum-straw ceiling ceiling boards (5– boards 10 cm in ∙ Thick brick wall (33– thickness) 45 cm) Sun shading ∙ Roof and/or ∙ Roof and/or ∙ Roof and window window overhang window overhang overhang, front or side ∙ Blinds or bamboo ∙ Shutter windows corridor screens ∙ Shutter windows Cooling Plants and small Plants and small Shade trees and plants with aquaria in aquaria in courtyards water surface in front and courtyards rear gardens traditional shophouse, French colonial shophouse and French colonial villa, all of which show some eco-features based on simple-but-efficient solutions as follows (Fig. 3.12 and Table 3.6). Morphologically, traditional and colonial row houses can be revitalised in modern row houses and self-built single-family houses as the two most popular townhouses whereas the colonial villa design experience is good enough to be directly transferred to new villas built in the twenty-first century.

3.4.2 Is There Another Way?

Another possibility does exit along with high-tech and low-tech housing concepts. It is worth trying to combine the strengths of the two design tendencies. In terms of achieving energy efﬁciency, if a high-tech housing concept seems to be accepted by high-income groups and a low-tech housing concept tends to serve low-income residents, a middle-of-the-road concept (or an in-between concept) should be targeted to middle-class people as a growing population group in most Vietnamese cities. There are many possibilities (scenarios) to propose, with various levels of high-tech and low-tech combination: 90 % + 10 %, 80 % + 20 %, 70 % + 30 %, 60 % + 40 %, 50 % + 50 %, 40 % + 60 %, 30 % + 70 %, 20 % + 80 % and 3 Achieving Energy Efﬁciency in Urban Residential Buildings … 65

10 % + 90 %. It is necessary to explore how the strengths of the two concepts can be maximised, corresponding to each level, and then to recommend a concrete and detailed guideline for each case. This will be very useful to all those who dream of living in a truly sustainable-but-affordable home.

3.5 Conclusion

Vietnam can make full use of the strengths, take the opportunities, self-correct the weaknesses and learn how to avoid the threats in developing energy efﬁciency in urban housing. Apart from low-tech design and high-tech design, a midpoint concept is appropriate and thus preferred, since the strengths of each solution will be combined. This possibility should be investigated for the future housing development.

References

Archdaily Journal of Architecture (2017) Eden villa project. https://www.archdaily.com/874585/ eden-villa-xyz-architects. Accessed 10 July 2018 Artemis Project Management Board (2017) Artemis apartment project. http://chungcuartemis.com/ thiet-ke-can-ho-the-artemis/. Accessed 25 June 2017 CapitaLand (2015) CapitaLand ranks top with most number of BCA Universal Design Mark Platinum Awards. https://www.capitaland.com/international/en/about-capitaland/newsroom/ news-releases/international/2015/may/nr-20150514-CapitaLand-ranks-top-with-most-number- of-BCA-Universal-Design-Mark-Platinum-Awards.html. Accessed 10 July 2018 European Chamber of Commerce in Vietnam Office (2017) Green Book—Renewable energy, waste, water, green buildings and smart cities: Best practices and solutions for smart and sustainable development in a new era for Vietnam and the EU, Hanoi, pp 10–11 Ministry of Construction (2011) National strategy for urban housing development towards 2020 with vision towards 2030, Hanoi, approved by the Prime Minister with Decision No 2127/ QD-TTg, Hanoi Ministry of Construction (2017) National Energy Efficiency Building Code, Hanoi Ministry of Construction (2018) Smart city as the solution to urban issues. http://www.moc.gov. vn/trang-chi-tiet/-/tin-chi-tiet/Z2jG/64/443323/xay-dung-do-thi-thong-minh-giai-phap-cho- cac-van-de-do-thi.html. Accessed 16 July 2018 Nation Online Map (2018) Vietnam administrative map. http://www.nationsonline.org/oneworld/ map/vietnam-administrative-map.htm. Accessed 10 July 2018 National Assembly12 (2010) Energy Saving and Energy Efficiency Law, Hanoi RSP Architects Vietnam Co Ltd (2018) Mulberry Lane project. http://www.rsp.vn/en/project/ detail/mulberry-lane-8.html. Accessed 10 July 2018 Singapore Building and Construction Authority (2014) Third Green Building Master plan, Singapore, p 3. https://www.bca.gov.sg/GreenMark/others/3rd_Green_Building_Masterplan. pdf. Accessed 10 July 2018 Statista (2017) Urbanisation in Vietnam. https://www.statista.com/statistics/444882/urbanization- in-vietnam/. Accessed 30 Nov 2017 UN Habitat Vietnam (2014) Vietnam housing sector, Hanoi, p 63 66 Q. M. Nguyen

Vietnam Association of Architects (2017) Criteria for National Green Architecture Awards. https:// kienviet.net/2014/10/07/infographic-5-tieu-chi-kien-truc-xanh-viet-nam/. Accessed 04 Oct 2017 Vietnam Electricity Corporation (2016) Press release—Abandoning nuclear power plant in Ninh Thuan province. https://evn.com.vn/d6/news/Thong-cao-bao-chi-ve-viec-dung-thuc-hien-Du- an-dien-hat-nhan-Ninh-Thuan-66-142-19153.aspx. Accessed 01 Dec 2016 Vietnam Green Building Council (2012). Green building evaluation in housing. http://vgbc.vn/en/ lotus-en/lotus-homes/, http://vgbc.vn/en/lotus-en/lotus-mfr/. Accessed 12 July 2018 Vietnam Green Building Council (2018) Green building database. http://vgbc.vn/en/vietnam- green-building-certification/. Accessed 10 July 2018 Vietnam Institute of Energy (2016) Vietnam Energy Statistics 2015, Hanoi, pp 29–31 Vietnam News (2017) Green is the way to go for housing in Vietnam, 2nd June 2017 Issue Vietnam News (2018) 485,000 kWh of electricity saved during Earth Hour 2018. https:// vietnamnews.vn/society/425090/485000-kwh-of-electricity-saved-during-earth-hour-2018. html#dZwvb3IjWS1rBcuD.97. Accessed 15 June 2018 Vo Trong Nghia Architect (2018) S-House-2 design. http://votrongnghia.com/projects/s-house-2/. Accessed 03 July 2018 Vu Phong Solar Company (2017) Solar energy in Vietnam. https://vuphong.vn/mien-bac-co-phu- hop-de-lap-dien-mat-troi-khong/. Accessed 12 Sept 2017 Wikipedia (2018) Vietnam Background map. https://en.wikipedia.org/wiki/Vietnam. Accessed 10 July 2018 World Atlas (2017) The biggest cities in Vietnam. https://www.worldatlas.com/articles/the- biggest-cities-in-vietnam.html. Accessed 24 Oct 2017 World Bank (2015) Affordable housing in Vietnam—A way forward, Hanoi, p 21 World Bank (2018) Vietnam’s urban population. https://data.worldbank.org/indicator/SP.URB. TOTL.IN.ZS. Accessed 22 May 2018 Worldometer (2018) Vietnam population. http://www.worldometers.info/world-population/ vietnam-population/. Accessed 06 June 2018 XYZ Design Saigon (2017) Eden villa project. http://xyzsaigon.vn/eden-villa/. Accessed 10 July 2018 Chapter 4 Recommendations for the Design of an Energy-Efficient and Indoor Comfortable Office Building in Vietnam

Ngo Hoang Ngoc Dung and Nguyen Trung Kien

Abstract The practice of energy efficiency to buildings requires a variety of interdisciplinary actions which are related to aspects of architecture and building services. To office buildings, it is more complicated for the fulfillment of energy efficiency and indoor comfort as such buildings’ designs are normally oriented in a way that creates mechanically air-conditioned spaces. In the context of Vietnam whose climate feature is classified as of humid tropical zone, the issue may become more serious and there is a need to look for new improvement in terms of architecture- and building service-related activities. The article provides an overview of actual situation of energy consumption and indoor conditions of office buildings in major cities of Vietnam, and, from the perspective of architectural and technical design, it gives ideas in aim of improving the energy efficiency and indoor comfort applied to the design concept of office buildings.

Keywords Building energy efﬁciency Á Building energy simulation Ofﬁce building design Á Climatic features of Vietnam

4.1 Problem Statement

The category of office building, together with high-rise apartment, have only emerged for around two decades in Vietnam, but the number of office projects has been accelerating in major cities namely Hanoi and Ho Chi Minh City in order to meet the large needs of work office for agencies and businesses. According to a report by CBRE Vietnam (CBRE Releases 2018), the total leasable floor area in the sector of office in Hanoi market is projected to reach 1.4 million sqm. by the end of

N. H. N. Dung (&) National University of Civil Engineering, Hanoi, Vietnam e-mail: [email protected] N. T. Kien Vilandco Company, Hanoi, Vietnam e-mail: [email protected]

© Springer Nature Singapore Pte Ltd. 2019 67 E. Motoasca et al. (eds.), Energy Sustainability in Built and Urban Environments, Energy, Environment, and Sustainability, https://doi.org/10.1007/978-981-13-3284-5_4 68 N. H. N. Dung and N. T. Kien

2018, with the presence of approximately 190 buildings, and at the same time, another report on the market of Ho Chi Minh tells a fact that total leasable office stock had achieved 1.7 million sqm. late 2017, and continue to grow in the next few years after the completion of projects which are now still under construction (Savills Vietnam 2017). The increasing number of office buildings inevitably results in a rapid rise in energy consumption of building sector. As an investigation into energy use of office building category conducted in five cities under the framework of Vietnam Clean Energy Program (VCEP), the largest part of energy consumption is for the operation of HVAC systems and office equipment. The HVAC accounts for the highest percentage of energy use of the buildings (statistically about 50%), while office equipment consume a smaller amount of electricity as it takes up around 24% of total value. (This information is provided by the USAID/Vietnam funded Vietnam Clean Energy Project implemented by Winrock International). The VCEP also presented energy use data by building applications in average EUI value (Energy use intensity, kWh m−2 year), and it indicates that normal building air conditioning EUI is of 50.66 kWh m−2 year, by far higher than that of lighting purpose (only around 19 kWh m−2 year) (This information is provided by the USAID/Vietnam funded Vietnam Clean Energy Project implemented by Winrock International). This breakdown of electricity consumption shows that energy use distribution of office buildings in Vietnam is not significantly distinguished to those built in other nations of various climatic conditions such as Australia where HVAC system of office building consumes an average amount of 50% of total site energy use (Parlour 2000) (Fig. 4.1). The responsibility of large amount of energy use by HVAC system in high-rise office buildings can be given to ‘close’ design characteristic which makes whole building isolated and totally environmentally regulated by air conditioner and mechanical ventilation. The survey data published by VCEP also reveals that the

Elevators LighƟng 11,1% 17,1%

Equipment Air 24,0% CondiƟoning 47,8%

Fig. 4.1 Share of consumed electricity of a typical office building in Vietnam (pie chart is created by authors based on data provided by the USAID/Vietnam funded Vietnam Clean Energy Project implemented by Winrock International) 4 Recommendations for the Design … 69 energy consumption, represented by BEI (building energy index) of the office buildings in Hanoi varies from 60 to 170 kWh m−2 year, while figure for the market of Ho Chi Minh city is in range from 90 to 220 kWh m−2 year (USAID Vietnam

Fig. 4.2 Energy consumption index of popular building categorizes in Hanoi (USAID Vietnam Clean Energy Program 2016)

Fig. 4.3 Energy consumption index of popular building categorizes in Ho Chi Minh city (USAID Vietnam Clean Energy Program 2016) 70 N. H. N. Dung and N. T. Kien

Outdoor air leakage Occupant Solar radiation heat

Lighting Heat transferred through envelop

Fig. 4.4 Distribution of cooling load toward a typical ofﬁce building in Hanoi (Tran Ngoc Quang 2015)

Clean Energy Program 2016). Office buildings are normally claimed to be a large energy consumer, raking behind only categorizes of hotel, retail and hospital buildings (Figs. 4.2, 4.3 and 4.4). Taken cooling load into a more detailed examination, solar radiation and heat transferred through building envelope are two main contributors in typical office buildings in Hanoi (Quang 2015), so it can be seen that building envelope plays crucial role in determining how much energy that building uses for cooling and heating purpose, though these loads are also influenced by technical system and occupancy schedule. Therefore, this chapter focuses its effort on determining optimized design for purpose of energy efficiency and indoor comfort of office buildings in Vietnam.

4.2 Analysis of Practical Situation and Scientiﬁc Basis for Recommendation

4.2.1 Overview of Actual Design Practice for Ofﬁce Buildings in Vietnam

Building envelope. Following the internationalized trend of contemporary architecture, a majority of modern office buildings in Vietnam are designed in a way that facilitates air conditioner use and, therefore, are intensively isolated by large amount of curtain and brick wall. A well-identified feature in the design of these buildings is glassy façade which intentionally provides sufficient daylight level and broadens occupants’ outlook. Turning back to years ago when insulation glass is hardly available, there was merely only choice to furnish external curtain wall with single clear glass which was already proved to be poorly energy efficient. Today, there is an increasing 4 Recommendations for the Design … 71 chance for insulation glasses that are more available in the market of Vietnam. The better accessibility of hi-tech glass, on the one hand, provides owners and architects with options to select glass material in accordance with aesthetic expectation of façade design, but, on the other hand, causes a fade in tropically strategic design approach due to a conception that high-performance glass can alternate climate-based architecture to move forward energy saving and indoor comfort. This also causes the emergence of a new trend in office building design in Vietnam which is not matched to local climatic condition, and inevitably, less energy efficient. Glass-coated envelope probably brings an apparently modern and flashy appearance to office buildings, but is unlikely to solve the energy-related issues to reduce greenhouse gas emission. In addition to the glazing, it is important to take the brick wall into consideration of building envelope performance as it affects the amount of heat transferred into indoor space. Other results of surveys showed a fact that only two among eighteen office buildings in survey are furnished with insulated walls whose U-value are below the maximum allowed. A similar situation is recorded in major cities of Vietnam where primary material for brick wall is brick-hollow with thickness of 220 m. Such kinds of wall possess heat conductivity which exceeds maximum value of heat transfer U-value (1.8 W m−2 K) as defined in the National Technical Code on Energy Efficient Buildings (QCVN 09: 2017/BXD) (Ministry of Contruction 2017). Only in a few buildings, external walls are structured with additional insolation layers like lightweight foam concrete with a thickness of 330 mm, or insulated block bricks (provided by the USAID/Vietnam funded Vietnam Clean Energy Project implemented by Winrock International). The total heat transfer value of these wall structure is measured to achieve 1.4–1.5 W m−2 K, which perfectly satisfies the requirement by building code. Unfortunately, most of the buildings in the examination are not insulated enveloped, and therefore do not meet the insulation requirements defined by QCVN 09: 2017/BXD.

4.2.2 Climatic Basis for Recommendation

Located in the zone bounded by the Equator and North tropic, Vietnam possesses a typical humid tropical climate which, however, uniquely characterized by cold winter resulted from Monsoon. Despite the fact that hot and humid climate is prevailing the whole country, there are still diversification between locations due to different topography. From the perspective of climatic classification, the country may be partitioned into two main regions as described below (Ministry of Construction: Vietnam Building Code Natural Physical and Climatic Data for Construction QCVN 02 2009): • The North region (from the latitude of 16° upward to the North part) is influenced by unique cold winter when mean air temperature falls down to between 10 and 15 °C; 72 N. H. N. Dung and N. T. Kien

• The South region (from the latitude of 16° downward) is prevailed by typical hot weather in a year round. There are no distinct cold and hot seasons, but significant difference in humidity is present, causing dry and rainy periods. Rainy season starts from May to October, and it turns dry and mild from November until next April. There is a great deal of sunny hours though the frequency of such state in the North and South regions are slightly varied. Number of sunny hours in the South exceeds 2000 h, while that of the North is below. Mean air temperature of the North normally is around 24 °C, but it hardly falls below, and may even achieves 28 °C in the South. As of tropical climate zone, the solar radiation intensity is always recorded high in all year round, and it reaches an average value of 586 kJ cm−2. Vietnam is characterized by high amount of humidity which often stays at 77– 87%, and it even goes up to maximum level during February and early of March in the Northeast and coastal Central area. It is important to mention an arid hot weather condition which lasts 10 to 30 days and occupies some mountainous parts of Central and North-west areas due to the operation of local Foehn breeze (Table 4.1). It can be seen that most parts of Vietnam’s territory have a hot season with high temperature, requiring heat insulation approach when it comes to the design practice. Despite the fact that several parts of the country are normally influenced by cold winter, it is demonstrated that air temperature during winter time may not be comparable to that of countries located in Europe and Arctic climate zone since the main cause of the local winter is extreme cold breeze generated by North-East Asian monsoon system, rather than existed low temperature background like European countries. Cold-proof techniques should only be taken into the design of buildings in some areas with very cold winters such as the Northeast, the Northwest and Highlands, and primary approach is simply preventing cold air flow coming into occupied spaces of the building. In the actual situation of Vietnam, high-rise office buildings are mainly located in urban areas where climates significantly differ from mountainous locations. Most of the cities are affected by hot weather condition identified by high solar radiation intensity while less influenced by cold winter; therefore, cooling by sun-shading and heat insulation approach are considered essential to the climatic design of office buildings. In addition, the high value of air temperature and large amount of water vapour in atmosphere causing high level of humidity is another factor to affect the indoor comfort of buildings. A study on psychrometric analysis for majority of Vietnamese cities by Duc Nguyen et al. (2005) shows a long time of the year when humidity exceeds comfort level despite the availability of climatic comfort time in all year round. Therefore, it is still recommended for office buildings to be air conditioned for the proper operation of office equipment (Fig. 4.5 and Table 4.2). Another study by Nguyen (2013) also reveals that thermal comfort zone can be extended by the practice of natural ventilation, but the presence of a number of 4 Recommendations for the Design … 73

Table 4.1 Summarizing of climatic features of subregions in Vietnam (Ministry of Contruction: Vietnam building code natural physical and climatic data for construction QCVN 02 2009) Region Subregion Location Features North IA Northwest and areas Most of the area suffers from extremely cold (I) bounded by Truong Son winter when air temperature may falls below mountain range 5 °C. It is also influenced by hot arid weather generated by local breeze in summer time, and may witnesses an up to 40 °C air temperature IB Mountainous Northeast The area possesses by far coldest winter due area to its high altitude compared to surrounding. The lowest temperature ever was recorded below 0 °C, but only appeared in several measure points on the mountain It may turn into a milder summer in comparison to North plain, but in valley locations, air temperature may rise to more than 40 °C. In mountainous locations, buildings are recommended to be cold-proof, rather than isolated from heat. Time when heating practice is needed may be extended to 120 days. Humidity value is always high in a year IC Northern plain It also suffers a cold winter but less extreme when compared to the region of IB. The variety of temperature and humidity is lower than those of IA and IB regions. The lowest temperature was caught to go below 5 °C, but the highest one may exceed 40 °C ID South of Northern region The highest temperature may reach 42 °C and North of Central region due to the presence of arid hot weather condition in summer. Heat insulation is an important approach, but cold-proofing should be taken into consideration during winter South IIA Coastal area of southern The area’s climate is generally a typical case (II) Central of tropic, except a small northern part of it is influenced by slightly cold breeze. Its lowest temperature is hardly below 10 °C, while the highest one normally exceeds 40 °C. Temperature varies slightly between days and nights, and there is no need for cold-proof practice IIB Tay Nguyen highland The weather condition significantly differs from regions whose altitudes are varied. While higher location may sometimes have cold weather pattern, rest parts of the whole area always suffer from hot summer, and require heat insulation applied to buildings IIC South region Its climate condition is typically tropical one with annual temperature of high value in all year round. Only two distinct dry and rainy seasons should be mentioned 74 N. H. N. Dung and N. T. Kien

Air-conditioning needed time

Fig. 4.5 Hourly plot weather data of Hanoi, Danang and Ho Chi Minh city on Building psychrometric chart at standard atmospheric pressure (101.325 kPa) (Nguyen 2013)

Table 4.2 Results of bioclimatic analysis for a number of cities of Vietnam (Duc Nguyen et al. 2005) City Very Cold Moderately Comfort Dry Humid Hot Very hot Very hot cold cold temperate temperate and humid and arid Ha Noi 0,60 8,60 18,00 44,60 0 23,40 4,50 0,30 0 Vinh 0,20 5,40 18,70 42,01 0 28,64 4,90 0,15 0 Da 0 0 4,53 85,42 0 8,85 1,20 0 0 Nang Nha 0 0 0 99,08 0 0,58 0,34 0 0 Trang Ho Chi 0 0 0,20 79,50 0 16,70 3,50 0,10 0 Minh Can 0 0 0 61,45 0 38,53 0,02 0 0 Tho

hours when humidity is beyond comfort level makes it unable to definitely depend on ventilation without air conditioning. These studies confirm the essence and relevance of air conditioning operation in office buildings of Vietnam. The overview of office buildings in Vietnam shows the inappropriateness in actual approach to the design of building envelope which increasingly relies on hi-performance glass material in façade while, at the same time, devalues climatic adaptation through passive sun-shading and heat insulation. There is, accordingly, a need for the removal of misconception and re-valuing of climatic-based design approach to envelope structures of office buildings in Vietnam.

4.3 Methodology

For a climatic-based design, the estimation of energy use and indoor comfort are valuable to provide stakeholders with comparison of performance between design options, and then help to determine the optimized approach. This practice is 4 Recommendations for the Design … 75 cooperatively supported by simulation software tools which give proper approximation of energy- and comfort-related parameters. In a practical project, simulation tools are important for designers during first steps of design to decide physical properties of building facades in terms of energy efficiency and indoor comfort, and simulation is an efficient and persuasive way to urge building owners to pursue the pathway of enhancing building performance. In this chapter, simulation-based method is used with a model of typical office building in Vietnam, and the results of simulation work are important basis for recommendation in the final section of the chapter. The main purpose of the simulation is to give a comparison in terms of efficiency and comfort between various options of glazing wall design for the building envelope. By doing so, there will be factors involving brick walls, occupancy and equipment to be kept remained and considered in ideal condition, while alternation of glass materials and the presence of shading device will be tested by simulation. Normally, the simulation work is carried out to the whole building to ensure the highest level of accuracy. However, this will require much energy and time, and become a real challenge to architecture team during conceptual design phase. Therefore, simulation practice in a way that is less effort-consuming is proposed, which selects a typical floor presenting the whole building layout to be examined in the simulation. The result of model simulation for typical floor will be used as the basis for efficiency assessment of the whole building. Although there might be a chance for certain errors due to the probable difference between floors, the uncertainty is acceptable in the very early stage of the project. Once a typical floor is selected, it will be partitioned into various thermal zones in accordance with solar heat gained of each façade before energy simulation starts. In the actual context of Vietnamese cities, office buildings are normally located in a land lot which has only one primary side accessible from urban streets, and the other sides are adjacent to surrounded constructions. The primary façade of the building is important factor that makes architectural aesthetic, while there is little requirement for secondary sides to be remarkably aesthetic. Therefore, the next step of simulation is to discover an optimized value of window to wall ratio (WWR) for the secondary facades which meet the target of solar heat gain reduction as well as enhancing daylight comfort. The found optimized value of WWR will be kept remained during the next simulation for the determination of best-matched design of primary façade. Various value of WWR on primary façade will be tested with simulation. The performance of each case will be evaluated upon aspects of energy efficiency, thermal comfort and lighting comfort, and cost-effective options will be drawn based on the analysis of simulation results. In particular, energy performance is shown by: • Cooling load and heating load which make senses in determining HVAC capacity and investment cost • Total site energy which combines energy use of all building elements and is useful when it comes to the estimation of operation cost. 76 N. H. N. Dung and N. T. Kien

Thermal comfort is assessed upon total number of unmet hours in a year. This indicator reveals how much time the setpoint of indoor climatic condition is maintained by HVAC operation, with permitted uncertainty is 1.1 °C. Lighting comfort is also taken into consideration by useful daylight illuminance (UDI) value which is deﬁned as the annual occurrence of illuminances across the work plane where all the illuminance values are within the range 100–2000 lux (Nabil and Mardaljevic 2005). The method combining analysis of thermal and lighting comfort for optimized envelope design was once mentioned and performed in a study for ofﬁce buildings of cold climate zone in Belgium by (Dartevelle et al. 2011) (Fig. 4.6). The principle of methodology is additionally illustrated by a diagram shown in Fig. 4.7.

Minimizing simulation work to the level of a typical floor

Thermal zone partitioning

Determining optimized WWR, Shading & Glazing type for secondary facades

Making comparison between options of WWR, Shading & Glazing type for primary facade

Energy efficiency/ Thermal comfort Daylighting comfort savings

Heating, Cooling load Radiance & Total Energy UDI, sDA & ASE Temperature Consumption

Fig. 4.6 Illustration of simulation principle 4 Recommendations for the Design … 77

North direction

Fig. 4.7 Sketch of models to be simulated

4.4 Model Simulation of a Typical Ofﬁce Building in Vietnam

4.4.1 Description of Simulation Model

The model represents a typical 20-storey office building in Vietnam whose form is in square-box with dimension of 30 Â 30 m. As mentioned in methodology section, the simulation work is performed to only a selected floor number 13, not to the whole building, for further assessment. Floor’s plan layout is arranged in a way that has gained popularity in office building category of Vietnam, in which technical

Zone 5- Office common area Zone 4- Office area

Zone 1- Technical core

Zone 3- Office area Zone 2- Office area

Fig. 4.8 Illustration of typical ﬂoor layout and its thermal zone division 78 N. H. N. Dung and N. T. Kien

Table 4.3 Functions and indoor setpoints of thermal zones No Zone name Zone type Input data 1 Zone 1 Lift & Stair (Unconditioned) Occupancy rate: Illumination: 110 lux Fresh air supply: 0.3 l/s m2 Heating setpoint: Cooling setpoint: Equipments: 2 Zone 2 Ofﬁce (Conditioned) Occupancy rate: 8 m2/person 3 Zone 3 Illumination: 300 lux Fresh air supply: 6.9 l/s person 4 Zone 4 Heating setpoint: 22 °C 5 Zone 5 Cooling setpoint: 25 °C Equipments: 11 W/m2

Table 4.4 Assumption of M&E operation No System name System type Specification 1 Lighting LED Luminaire type: recessed LPD: 6 W/m2 No lighting control 2 HVAC VRV (with AHU) Capacity: autosize COP: 3.3 (for heating and cooling) No heat recovery equipment 3 DHW Electric resistant Capacity: Autosize Efficiency: 0.9 core is well laid to behind edge of the building and usable office area is located around the core and along the other edges. Plan layout is illustrated in Fig. 4.8. The simulation work will be run with software tool named DesignBuilder v5.4. Representing a real medium-size office building, the model is assumed to have main access from an urban street, while its other sides area next to neighbouring constructions. The primary façade is right southward oriented. Information of functions and setpoints for each thermal zone are summarized in Table 4.3, and assumption data of M&E system is shown in Table 4.4.

4.4.2 Simulation in Search for Optimized Value of WWR on Secondary Facades

The first simulation is done with variation in value of WWR on secondary facades to determine what optimized WWR on these sides should be. Various values of WWR as 30, 40, 50, 60, 70, 80 and 90% will be applied to three secondary facades, in one case of no shading, and two other cases of horizontal and box-type shading devices. The overhang of shade in any case is 0.5 m. Another assumption is that 4 Recommendations for the Design … 79 glass type is clear single with 3-mm thickness whose SHGC and VLT are of 0.819 and 0.881, respectively. In this simulation, the model is structured with concrete mansory unit whose U-value is 1.204 W m−2 K. This is a well-insulated brick, and definitely compliant to insulation requirement as defined by regulation QCVN 09:2017/BXD (Ministry of Construction 2017). Roof structure consists of 4 layers with U-value of 1.589 W m−2 K. All data of building elements are shown in Table 4.5 (Figs. 4.9, 4.10, 4.11 and 4.12). The indicators for determination of optimized WWR are value of heating and cooling loads. Simulation results are illustrated by following diagrams. It can be seen from the chart that cooling and heating load will be minimized in case the value of WWR is 30% (illustrated by the blue color lines in the four charts), regardless whether shading is provided. The presence of shading makes the decrease in both cooling and heating loads, and how much these loads can be reduced depends on the sufficiency that shading can achieve.

Table 4.5 Input data of building elements No Structure Orientation Speciﬁcation 1 Wall All 250 mm CMU (Plaster/CMU/Plaster) U-value = 1.204 W/m2 K 2 Roof – Ceramic tiles/Concrete slab (0.15)/Air 3 Floor – gap (0.3)/Gypsum board U-value = 1.589 W/m2 K 4 Glazing East, West and North Single clear with shading (combine 0.5 m horizontal and vertical shading) South Varies with and without shading

No shading 0.5m Overhang overhang + sidefins

Fig. 4.9 The dependence of cooling load on the variation of WWR value and shading types in zone 2 80 N. H. N. Dung and N. T. Kien

No shading 0.5m Overhang overhang + sidefins

Fig. 4.10 The dependence of heating load on the variation of WWR value and shading types in zone 2

No shading 0.5m Overhang +

overhang sidefins

Fig. 4.11 The dependence of cooling load on the variation of WWR value and shading types in zone 4

To test the dependence of lighting comfort on the amount of window which may have signiﬁcant inﬂuence on human behaviour, the indicator of annual sun expose (ASE) will be simulated and used as the basis for the assessment. The ASE is given to describe how much of space is affected by more than 250 h of direct illumination (higher than 1000 lux) per year (Barbara Gherri 2015), which can cause visual discomfort (glare) or increase cooling loads. Simulation results of ASE in the two cases where WWR is of 30 and 90% on secondary facades (North-, East- and West-oriented) are illustrated in Table 4.6. 4 Recommendations for the Design … 81

No shading 0.5m Overhang + overhang sidefins

Fig. 4.12 The dependence of heating load on the variation of WWR value and shading types in zone 4

Table 4.6 Result of ASE calculation for deﬁned zones while WWR value and shading type vary WWR (%) Louvre shading Zone Area (m2) ASE area ASE area in area (m2) in area (%) 30 Yes 2 214.926 63.826 29.7 4 214.926 66.31 30.85 30 No 2 214.926 65.138 30.31 4 214.926 67.97 31.62 90 Yes 2 214.926 46.505 21.64 4 214.926 65.58 47.538 90 No 2 214.926 50.461 23.48 4 214.926 52.31 24.34

It is clear that shading efficiently helps to reduce space affected by exceeding 1000 lux illuminance in both zones (down to 29.7 and 30.85%, compared to 30.31 and 31.62% when no shading is available) if the proportion of window remains at 30%. Even if that amount rises up to 90%, shading still works, but the value of ASE is moderately increased by about up to 1.5 times. The results can lead to a conclusion that a reasonably small value of WWR (30%) may only produce sufficient daylight comfort when coupling with efficient shading by reducing indicator of ASE. Therefore, optimized secondary facades will be set with WWR of 30% and with box-type shading devices as input data for further simulation to determine best-matched properties of primary (South-oriented) façade. 82 N. H. N. Dung and N. T. Kien

4.4.3 Simulation for Optimized Primary (South-Oriented) Façade

In practice, the WWR for the main façade is often required to be higher than the secondary façades to create aesthetic highlights for the building. However, there are cases where the owner does not require a large ratio of window, so the WWR of 30% can still be possible. In order to assess the impact of WWR on the main façade to the energy efficiency, thermal and daylight comfort will be simulated with various WWR values of 30, 50, 70 and 90%, respectively. Heating load, cooling load and total site energy, which is the total energy consumed by all systems and electrical appliances in the building (HVAC, lighting, domestic hot water—DHW) are taken into consideration. Furthermore, thermal comfort is indicated by the number of hours when comfort target set by the air conditioner system (as stated in the setpoint of thermal zones) is unable to meet (Fig. 4.13). Based on the Sun path diagram of Hanoi (latitude of 21oN), and assuming the south face should be completely sunscreened from 8.00 to 17.00 on a daily basis, best-suited typology of shade can be drawn as of horizontal form whose shading mask is overriding illustrated to shade-needed zone in Fig. 4.14. The overhang of shade is supposed to be in the range of 1.5–2 m long for sufficient shading capacity, but is considered a misbegotten form in terms of structure and aesthetic. Therefore, horizontal shade should be alternated with louvres with both overhang and spacing of 0.5 m to maintain shading efficiency and to be more technically suited.

Fig. 4.13 Shade-needed zones and mask of shading devices 4 Recommendations for the Design … 83

Fig. 4.14 Louvers with overhang of 0.5 m (on the right) are more technically possible compared to horizontal shade with overhang of 2 m (on the left)

In addition to testing efﬁciency with variation of WWR, different glass options will also be experimented as an effort to discover a combining way of WWR, shading typology and glass category for optimized efﬁciency and indoor comfort. The types of glass tested in the simulation are shown in Table 4.7.

Table 4.7 Types of glass in simulation test No Glazing type SHGC (g-value) VLT U-value 1 Single clear 6 mm 0.819 0.881 5.778 2 Single low-E 6 mm 0.72 0.811 3.779 3 Double clear 6 mm/13 mm air 0.703 0.781 2.665 4 Double low-E clear 6 mm/13 mm argon 0.373 0.444 1.493

Table 4.8 Southward WWR of 30% with no presence of shading Heating Cooling Total site Unmet UDI load (kWh) load (kWh) energy (kWh) hours Single clear 6 mm 563.60 1853.57 1961942.22 224 50.42 Single low-E 6 mm 552.64 1842.55 1964013.53 223 49.93 Double clear 6 mm/ 547.49 1836.09 1955673.88 213 49.49 13 mm air Double low-E clear 538.42 1816.78 1925459.75 183 46.55 6 mm/13 mm argon 84 N. H. N. Dung and N. T. Kien

Table 4.9 Southward WWR of 30% with presence of shading Heating Cooling Total site Unmet UDI load (kWh) load (kWh) energy (kWh) hours Single clear 6 mm 563.60 1842.31 1924307.22 177 44.90 Single low-E 6 mm 552.64 1839.62 1926438.95 176 45.83 Double clear 6 mm/ 547.49 1833.47 1921533.14 176 44.68 13 mm air Double low-E clear 538.42 1814.90 1905022.92 173 43.15 6 mm/13 mm argon

Simulation results. Following tables show simulation results with variation of WWR, shading types and glass use on primary façade (Tables 4.8 and 4.9). Simulation results for the WWR of 30% also reveal that what glass type is used does not have significant impact on loads. In case shading is absence, double low-E glass helps to reduce only 1.9% in cooling load (1816.78 kWh) when compared to the case that it is replaced by single clear glass (1853.57 kWh). A similar outcome is drawn with the presence of louvre-shading. The effectiveness of sun-shading is not only in reducing cooling load but also in decreasing the number of hours in which comfort is not achieved. Specifically, comfort level generated by single clear glass with shading (only 177 unmet hours) is even higher than that when alternated with double low-E but no sunscreen attached (183 unmet hours). Louvres shading do not much affect the UDI index, as the maximum difference between two cases of with or without sunshade is only 7% (UDI is 50.42% for single clear glass without sunscreen, and is 43.15% for double low-E). The figures demonstrate the harmlessness of shading to the quality of daylighting. By increasing window ratio to 50%, the demand for cooling, heating and power consumption also increases, and leads to the rise in number of met hours. However, sufficient sun-shading can be used to provide higher thermal comfort with the presence of only single clear glass in comparison with that generated by double low-E but non-sunscreen attached. A similar effectiveness of sun-shading in terms of loads and comforts is also demonstrated when WWR value is adjusted to 70%.

Table 4.10 Southward WWR of 50% with no presence of shading Heating Cooling Total site Unmet UDI load (kWh) load (kWh) energy (kWh) hours Single clear 6 mm 577.74 1883.42 2000450.67 349 54.53 Single low-E 6 mm 560.00 1881.30 2009904.95 394 55.09 Double clear 6 mm/ 553.38 1871.88 1995260.14 363 54.80 13 mm air Double low-E clear 538.94 1834.03 1949967.36 208 52.27 6 mm/13 mm argon 4 Recommendations for the Design … 85

The simulation results for WWR of 50 and 70% are presented in Tables numbered 4.10, 4.11, 4.12 and 4.13. For the case where WWR rises to 90%, demand for cooling, heating and power consumption is increased to the highest level, and thermal comfort is affected in the most negative way by maximizing number of unmet hours. In this circumstance, effectiveness of low-E glass is most noticeably recognized, and significantly more efficient than low insulated glass. The most common case is the smaller WWR. However, the simulated results in Table 4.14 show that low-E glasses exhibit significantly lower thermal and thermal efficiency than glass in a single layer. Whereas, if the amount of window on the southern façade is of a much smaller value, energy savings generated by low-E is only slightly differentiated from that resulted from single clear glass. For example, if double low-E glass is applied to the

Table 4.11 Southward WWR of 50% with presence of louvers Heating Cooling Total site Unmet UDI load (kWh) load (kWh) energy (kWh) hours Single clear 6 mm 577.74 1879.83 1939858.67 185 52.00 Single low-E 6 mm 560.00 1875.98 1945754.71 186 49.99 Double clear 6 mm/ 553.38 1867.14 1937889.82 185 51.72 13 mm air Double low-E clear 538.94 1830.66 1913870.81 175 52.36 6 mm/13 mm argon

Table 4.12 Southward WWR of 70% with no presence of shading Heating Cooling Total site Unmet UDI load (kWh) load (kWh) energy (kWh) hours Single clear 6 mm 590.37 1910.06 2030828.39 499 58.07 Single low-E 6 mm 566.60 1908.14 2051394.22 534 57.59 Double clear 6 mm/ 558.88 1895.94 2030291.04 491 57.64 13 mm air Double low-E clear 539.39 1849.48 1973661.07 286 57.67 6 mm/13 mm argon

Table 4.13 Southward WWR of 70% with the presence of louvers Heating Cooling Total site Unmet UDI load (kWh) load (kWh) energy (kWh) hours Single clear 6 mm 590.37 1905.48 1954735.92 174 56.00 Single low-E 6 mm 566.60 1901.27 1967257.47 178 56.15 Double clear 6 mm/ 558.88 1889.99 1955164.61 183 57.37 13 mm air Double low-E clear 6 539.39 1845.07 1925467.46 183 56.82 mm/13 mm argon 86 N. H. N. Dung and N. T. Kien

Table 4.14 Southward WWR of 90% with no presence of louvers Heating Cooling Total site Unmet UDI load (kWh) load (kWh) energy (kWh) hours Single clear 6 mm 601.71 1934.55 2054129.33 617 62.24 Single low-E 6 mm 572.67 1933.38 2088454.08 678 59.46 Double clear 6 mm/ 564.00 1898.85 2008734.47 418 60.10 13 mm air Double low-E clear 539.78 1873.13 1996217.55 377 58.63 6 mm/13 mm argon

South façade whose windows take up to 90%, its corresponded cooling load falls by 3% to 1898.85 kWh when compared to that generated by single clear glass (1934.55 kWh), and number of unmet hours contemporaneously decreases by 39% to 377 h. Similarly to cases where the smaller WWR value is applied, shading still provides efficiency and indoor comfort even if window ratio is increased to 90%. More noticeably, the number of unmet hour when southward façade is furnished with only single clear glass with shading is (161 h) by far reduced by 74% if no sunscreen is attached (617 h). This also generates a total energy consumption of 13% lower than in the case that double low-E is applied but no sunscreen is attached. Through simulation, it is possible to blame energy loss, inefficiency and discomfort in summer for large amount of glazing wall though it may generate an increase in UDI. In condition of North region of Vietnam where cold winter lasts for at three months (taking up to 8.6% of the year time Ministry of Construction: Vietnam Building Code Natural Physical & Climatic Data for Construction QCVN 02 2009), glazing wall is also a contributor to the heat loss for heating operation. A study on the influence of glazing use upon energy efficiency by Nguyen Van Muon (Nguyen Van Muon 2015) demonstrates a fact that in all cases of climate condition, large window wall holds main responsibility for energy loss of HVAC system. Low-E glass is useful to reduce loads in both summer and winter, and at the same time reduces the number of unmet hours, but its effectiveness is strongly noticeable in case large amount of glass is incorporated in building façade (WWR of such 70 and 90%). In the case of relatively small value of WWR (30%), the influence of glass on loads and energy use is negligible. Solar shading is the key factor to achieve high performance and indoor comfort regardless how insulation level of glass, which brings an opportunity to replace costly hi-tech glass with more economical one for better savings on investment cost. In a tropical climate condition characterized by high intensity of solar radiation of Vietnam, envelope shading is extremely important and brings about clear opportunity of energy saving and reducing discomfort time. Unlike misconception by a group of project developers, sun-shading does not affect illuminance comfort 4 Recommendations for the Design … 87 as UDI values still remain reasonable even though it is slightly lower than that in replace hi-tech insulation glass with conventional single glazing while still ensure energy performance.

4.5 Recommendation

4.5.1 Recommendation for Energy-Efﬁcient Building Envelope

Building envelope plays a crucial key to energy performance and indoor comfort of a building. Under climate condition of Vietnam, there is always a need for intensive insulation to envelop structure of all building categories, and especially important to high-rise office buildings. As a useful approach, reducing solar heat transferred through curtain walls will help to lessen cooling loads on HVAC system, and therefore save more on energy cost. Additionally, insulation for brick walls is also essential for a more efficient design of building envelope. Curtain wall. It is strongly recommended that secondary facades which are not mandated to be aesthetically satisfied that the value of WWR should not be set exceed 30% to ensure small cooling load on HVAC and sufficient indoor comfort. However, the value of WWR for primary façade can be differentiated from secondary ones, letting the index reach up to 90% as a way to highlight building’s architectural features, but passive shading by louvres or best-matched shading devices in combination with low-E high-performance glass is always advisable for better energy savings and indoor comfort. In case of modest investment cost, there might be a possibility to replace hi-tech glass with single clear glazing wall, but sufficient shading is always recommended. Low-E glass is obviously effective in terms of insulation, but large proportion of low-E glass without shading on building façade will cause heat loss through envelope as well as thermal and daylight discomfort by increasing number of unmet hours and level of discomfort glare. Incorporating a large amount of glazing wall into building façade without shading is never an effective way to enhance building performance and indoor comfort under the condition of tropical climate of Vietnam. In case that glazing window takes up a rather small part of the façade (WWR = 30–40%), it might be more cost-effective to replace low-E with double, or even single glass while, at the same time, still remain energy efficiency and indoor comfort if shading devices is incorporated. Brick wall. Part of brick wall is also factor that affects energy consumption, and therefore should be well-structured to reduce its heat transferred value (U-value). It is advisable to select light-weighted porous concrete unit or multilayered wall structure whose thermal resistant is sufficient to meet the insulation requirement for external walls defined by the National regulation on energy-efficient buildings (QCVN 09:2017/BXD). 88 N. H. N. Dung and N. T. Kien

4.5.2 Additional Technical Measures to Enhance Building Energy Performance and Indoor Comfort

In addition to building envelope, the M&E equipment also plays an important part in the determination in energy end use of building. Upon the existing model which is shown in Table 4.15, extra simulation is performed with alternation of HVAC and lighting system whose COP and LPD vary. The result of simulation work with various values of COP and LPD reveals that saving level can be improved when more efficient HVAC (high value of COP) and lighting system (lower value of LPD) are combined with architectural approach (Tables 4.16 and 4.17). Another recommendation in an effort to take advantage of climatic benefits and reduce energy use of air conditioner is to operate the building in a mixed mode. This means the air conditioning system may be automatically adjusted in accordance with outside atmosphere state so that only AHU is in operation (no cooling mode) if outside air falls below setpoint which is already mentioned in the section of input data description. It is possible for the idea to be turned into reality as majority of Vietnamese cities have at least about 40% of the time (Duc Nguyen et al. 2005) in a year when the weather condition is in comfort situation. The simulation results show the total site energy use may decrease by 10.88% to approximately 1755695.01 kWh when only mix-mode is operated, and if all suggested efficiency measures are taken into buildings, saving levels will go up to 36%, making a fall in energy use to only 1245193.35 kWh.

Table 4.15 Southward WWR of 90% with the presence of louvers Heating Cooling Total site Unmet UDI load (kWh) load (kWh) energy (kWh) hours Single clear 6 mm 601.71 1929.10 1970082.09 161 58.16 Single low-E 6 mm 572.67 1925.03 1990947.14 170 57.30 Double clear 6 mm/ 564.00 1911.27 1973919.08 174 58.65 13 mm air Double low-E clear 539.78 1858.68 1939468.15 185 59.24 6 mm/13 mm argon

Table 4.16 The dependence of energy savings on COP COP 3.3 3.6 3.9 4.2 4.5 4.8 Total site 1970082.09 1909034.61 1857379.05 1813102.86 1774730.16 1741154.05 energy Savings (%) 0.00 −3.10 −5.72 −7.97 −9.92 −11.62 4 Recommendations for the Design … 89

Table 4.17 The dependence of energy savings on LPD LPD 6 4.5 3 Total energy 1970082.09 1873395.46 1777362.52 Savings (%) 0.00 −4.91 −9.78 With lighting control 1710778.73 1677236.88 1644107.34 Savings (%) −13.16 −14.86 −16.55

These tests also demonstrate a great signiﬁcance of coupling climatic-based design of building envelope with high-performance equipment as an attempt to enhance building energy efﬁciency and indoor comfort under climatic condition of Vietnam.

4.6 Conclusion

The physical feature of office building envelope is a key factor to determine its energy use and saving, but how the glazing part is treated is even more important. The simulation results prove that large area of sun-exposed glazing wall on office building façade is the main contributor to energy loss and energy inefficiency in the condition of Vietnamese tropical climate. Therefore, high-rise office building is always recommended to be well-shaded by sufficient external shading which may be in form of louvres as well as sidefins, though high-performance glass units (normally coated low-E) can be partly used to enhance capacity of solar heat reduction. However, as demonstrated by simulation results, insulation glass does not always provide building with effective insulation, and therefore cannot alternate external shading for more improved insulation and energy performance of building. A sun-shaded building envelope is also an identity of Vietnamese tropical architecture which was once spontaneously internationalized and now needs reviving. In addition to the energy-efficient building envelope, taking advantages of climatic benefits through practice of mix-mode operation is possible and useful to the enhancement of indoor comfort in the climatic condition of Vietnam. This is even more beneficial to occupants’ health when fresh air flow is possibly increased and likely sick building syndrome may be minimized. For a climatic design of an office building, simulation tool is useful for the estimation of building energy performance, and it may give designers first comparison between options of concept design and then help in shortlisting best-matched approaches. In early stage of the design, the important point is to determine which envelope option can be proceeded based on energy aspect while primary function and plan layout is normally duplicated to a large number of floors in the building, so simulation can be performed with only a typical floor rather than the whole building in aim of time and effort saving. Efficiency-related results of the whole building can be deduced from those of typical floor with reasonable modification due to the similarities between building floors. 90 N. H. N. Dung and N. T. Kien

References

CBRE Releases (2018) Quarterly Report Highlights Hanoi Market. http://www.cbrevietnam.com/ Vietnam-Property/pressrelease/cbre-releases-q2-2018-quarterly-report-highlights-hanoi-market Dartevelle O, Deltour J, Bodart M (2011) Coupling thermal and daylighting dynamic simulations for an optimized solar screen control in passive office buildings Duc Nguyen P et al (2005) A ministerial research project on bio-climatic database for construction practice in Vietnam Gherri B (2015) Assessment of daylight performance in buildings. WITpress Ministry of Construction (2017) National technical regulation on energy efficiency buildings QCVN 09:2017/BXD Ministry of Construction: Vietnam Building Code Natural Physical & Climatic Data for Construction QCVN 02:2009/BXD Nabil A, Mardaljevic J (2005) Useful daylight illuminance: a new paradigm for assessing daylight in buildings Nguyen AT (2013) Doctoral thesis—Sustainable housing in vietnam: climate responsive design strategies to optimize thermal comfort Quang TN (2015) University Research Project—Influence of Building Services System on Building Architecture and Structure Parlour, R.P.: Building services: a guide to integrated design: engineering for architects. Integral Publishing, Pymble, NSW (2000) Savills Vietnam (2017) Reports on HCMC real estate market in Q4/2017. http://www.savills.com. vn/_news/article/31256/157903-0/1/2018/savills-vietnam-reports-on-hcmc-real-estate-market- q4-2017 USAID Vietnam Clean Energy Program (2016) Annual report FY 2016, 1 Oct 2015–30 Sept 2016 van Muon N (2015) Building with large glazed areas—energy killer Part II Energy Sustainability Strategies Chapter 5 Linking Neighborhoods into Sustainable Energy Systems

A. T. D. Perera, Silvia Coccolo, Pietro Florio, Vahid M. Nik, Dasaraden Mauree and Jean-Louis Scartezzini

Abstract Improving the energy efficiency and sustainability in the urban sector plays a vital role in the energy transition. Hence, it is important to consider promising ways to design sustainable urban energy hubs linking neighborhoods into energy systems. Improving the efficiency and sustainability of urban energy infrastructure is a process with multiple steps. This chapter presents the workflow that is required to be followed in this process. A brief overview about the methods that can be used to consider urban climate, urban simulation, and energy system design are presented in this chapter highlighting the crosslinks among these topics. Finally, the chapter presents the research gaps and promising areas to conduct future research.

Keywords Energy systems Á Urban energy modeling Á Microclimate Climate change

A. T. D. Perera (&) Á S. Coccolo Á P. Florio Á D. Mauree Á J.-L. Scartezzini Solar Energy and Building Physics Laboratory (LESO-PB), Ecole Polytechnique Fédérale de Lausanne (EPFL), 1015 Lausanne, Switzerland e-mail: dasun.perera@epﬂ.ch V. M. Nik Division of Building Physics, Department of Building and Environmental Technology, Lund University, 22363 Lund, Sweden V. M. Nik Division of Building Technology, Department of Civil and Environmental Engineering, Chalmers University of Technology, 41296 Gothenburg, Sweden V. M. Nik Institute for Future Environments, Queensland University of Technology, Garden Point Campus, 2 George Street, Brisbane, QLD 4000, Australia

© Springer Nature Singapore Pte Ltd. 2019 93 E. Motoasca et al. (eds.), Energy Sustainability in Built and Urban Environments, Energy, Environment, and Sustainability, https://doi.org/10.1007/978-981-13-3284-5_5 94 A.T.D. Perera et al.

5.1 Introduction

Over 50% of the world population now live in urban areas and this figure is expected to rise to above 70% by 2050 (ICLEI 2009) when cities will have to accommodate an additional 2.5 billion inhabitants. Energy sustainability at the urban scale is therefore vital. Building stocks will need to be linked to sustainable energy generation sources to a much greater extent, which makes it important to have a detailed understanding of distributed energy demands, potential for on-site power generation using renewable energy technologies, and optimum energy mix between different energy technologies and their interaction within multi-energy networks (refer to Fig. 5.1). Once the energy infrastructure has been designed or upgraded, it is furthermore important to operate it in an optimum way while minimizing the cost and guaranteeing the stability and reliability of the of energy services provided. Each of the aforementioned points corresponds to a separate research problem as shown using different modules in Fig. 5.1. The most challenging part is to maintain the information flow through these boxes in order to arrive at truly sustainable solutions. This chapter presents the “big picture” on how to manage the workflow when combining several models through a computational platform. Developing a computational platform linking different aspects related to urban energy system design and urban planning is a challenging task. We limit the scope of this chapter, to consider only the design process of distributed energy systems and urban energy simulation without further discussing dispatch optimization and

Fig. 5.1 Linking neighborhoods into sustainable energy systems 5 Linking Neighborhoods into Sustainable Energy Systems 95 energy networks. This chapter elaborates in detail on how to develop a computational platform combining: • future climate conditions and urban climate using regional and urban climate models (Sect. 5.1.1); • assessment of the energy demand of a building stock considering the urban climate and the interaction among buildings (Sect. 5.1.2); • design and assessment of complex urban energy systems considering the wind speed and solar irradiation (Sects. 5.1.3 and 5.1.4); • research gaps in present state are promising pathways to address the present research gaps are discussed in Sect. 5.1.6.

5.1.1 Bringing up Future Climate Conditions for Energy System Design

To make cities sustainable, it is essential to understand and to model urban meta- bolism, thereby interconnecting the urban energy fluxes that keep a city alive. Urban Building Energy Modeling (UBEM) is a nascent field, based on the application of physical models (heat and mass transfer) inside and outside a group of buildings. UBEM predicts buildings’ energy performance, operation of energy systems as well as indoor and outdoor environmental conditions (Reinhart and Davila 2016). Due to the complexity of the urban environment, several data are required for this purpose. These can be classified into the following three categories: • Climatic data: (i) meteorological data, for a typical year (Remund et al. 2015)or by monitoring and (ii) microclimatic data, as a function of city design. • Outdoor environment: (i) greening (grass and tree), (ii) water bodies, and (iii) albedo and thermal properties of the urban materials. • Buildings: (i) physical properties of the envelope (e.g., U-value of walls and windows, etc.), (ii) function of the building (appliances, occupant density, and profile), and (iii) renovation scenarios • Energy system: (i) available technologies for power generation and storage, (ii) techno-economic data for renewable energy components, and (iii) regulations for grid integration and building related renewable energy integration.

5.1.1.1 Considering the Urban Complexities

Understanding the urban climate conditions plays a vital role when considering both demand and generation. It is already well known that the presence of buildings and other artificial surfaces significantly influences the weather patterns in urban 96 A.T.D. Perera et al. areas. For example, the modification of the wind speed and the trapping of heat in cities were described in detail by Oke (1967, 1982) as the Urban Heat Island (UHI) phenomenon. The specific processes driving the urban climate are mostly due to the following: • Trapping of heat (inter-reflections, albedo). • Modification of wind patterns (drag and shear forces). • Lack of evapotranspiration. The study of the climatic conditions began with the need to understand air pollutant dispersion in urban areas but has been extended to include outdoor human comfort, energy system design, and building energy use in the recent years. The energy demand of buildings has an obvious relationship with the meteorological conditions (Kohler et al. 2016). Multiple studies have shown the correlation between the two (Ashie et al. 1999; Wang and Chen 2014). There are however some specificities when looking at the energy consumption of buildings in urban areas. At midlatitudes, the UHI tends to decrease the energy consumption for heating during the winter time while increasing the need for cooling in the summer time. The increase in the summer time will be further exacerbated in the future due to climate change (Mauree et al. 2018).

5.1.1.2 Considering Future Climate and Extreme Climate Scenarios

Climate projections show changes in average conditions of climate and in its variability, including changes in the frequency and magnitude of extreme events, which will be more frequent and stronger in the future (Field et al. 2012). The sustainability and resilience of energy systems and urban areas require estimating the probable future conditions and adapting to them. A large amount of work exists on assessing the impacts of climate change on buildings as the main energy users in cities and urban areas, focusing on different categories of buildings (e.g., de Wilde and Coley 2012; Kalvelage et al. 2014; Shibuya and Croxford 2016), thermal comfort and indoor conditions (e.g., Alves et al. 2016; Barbosa et al. 2015; Fisk 2015), building envelope and retroﬁtting strategies (e.g., Chow et al. 2013; Nik et al. 2012a; Karimpour et al. 2015), energy saving potentials, and building components (Shibuya and Croxford 2016; Nik et al. 2012a; Pakkala et al. 2014). Beyond energy demand, climate change can also affect energy generation, especially renewable generation such as wind (Pryor et al. 2006; Seljom et al. 2011), hydropower (Kao et al. 2015), and solar energy (Fant et al. 2016). Planning for climate change adaptation is complicated since it is difﬁcult to predict the expected degree and pace of warming (The Global Risks Report 2016). Impact assessment of climate change is usually performed by means of the climate data generated by Global Climate Models (GCMs), which simulate future climate conditions for the spatial resolution of 100–300 km2 (Meehl et al. 2007). Direct use of GCMs’ outputs is not recommended due to recognized biases (Fowler et al. 5 Linking Neighborhoods into Sustainable Energy Systems 97

2007; Prudhomme et al. 2010) and their coarse resolution (outputs cannot be considered as weather). Therefore, GCMs’ outputs should be downscaled to finer spatial (and temporal) resolutions, using statistical or dynamical downscaling methods. One well-known statistical technique is morphing (Belcher et al. 2005) with the drawback of reflecting only changes in the average weather conditions, underestimating climate variations and neglecting extreme conditions. Dynamic downscaling of GCMs by means of Regional Climate Models (RCMs) has the advantage of generating physically consistent data sets across different variables (CORDEX 2016; Giorgi 2006; Samuelsson et al. 2011). The weather data used in this work has been downscaled using RCA4; the fourth generation of the Rossby Centre regional climate model (Samuelsson et al. 2015). RCA4 downscaled four driving models to the spatial resolution of 12.5 km: CNRM-CERFACS-CNRM-CM5, ICHEC-EC-EARTH, IPSL-IPSL-CM5A-MR, and MPI-M-MPI-ESM-LR (which are called CNRM, ICHEC, IPSL, and MPI hereafter, respectively). The driving models are forced by two Representative Concentration Pathways (RCPs) (Giorgetta et al. 2013); the first two are forced by RCP4.5 and RCP8.5 and the other two by RCP8.5, resulting in six future climate scenarios in total. Two major challenges in the impact assessment of climate change are dealing with climate uncertainties and large data sets (e.g., Nik et al. 2012b; Nik 2010). It is not possible to rely on short time spans when working with future climate scenarios; periods of 20–30 years should be considered. Moreover, there are different uncertainties that affect the simulated climate data such as the selected GCM, RCM, emissions scenario, and the spatial resolution (Nik et al. 2012b). Hence, a valid impact assessment should consider several climate scenarios and not just a few (IPCC 2007; Kjellström et al. 2011; Christensen et al. 2010; Kershaw et al. 2011). This means that to have a valid energy assessment, several long-term weather data sets should be used, or statistically representative data sets should be generated, considering extreme climatic conditions. Furthermore, it is important to consider the impact of extreme climate on urban microclimate (Fig. 5.2).

5.1.1.3 Impact of Climate Change and Extreme Climate Events on Urban Climate

Several tools have been developed recently to consider the impact of urban climate on the energy demand of buildings and energy generation (especially considering the wind power generation). These tools have been coupled with meteorological models (Krpo et al. 2010; Pigeon et al. 2014; Salamanca and Martilli 2010). However, considerable computational time is required (Martilli 2007) and the horizontal and vertical resolution of these models remains very coarse and hence does not provide the boundary conditions appropriate for building energy simulation models (Garuma 2017; Mauree et al. 2018). Some studies have tried to reduce this gap with the development of models that can provide a link between regional climate models and micro-scale models (Bueno et al. 2013; Mauree et al. 2015, 98 A.T.D. Perera et al.

Fig. 5.2 Steps followed to consider climate uncertainty and extreme climate events

2017). These connecting models can be used to transform data obtained either from meteorological models or from monitored stations or typical years to a more localized dataset that would take into account the multiple processes in urban areas. Their inclusion has proved to improve accuracy in the simulation of the energy demand (Mauree et al. 2017) and to determine more precisely the peak energy demand, which can signiﬁcantly affect the energy system sizing (Perera et al. 2018). The methods have been used to calculate the energy demand at both the city and the district scale. Both the local meteorological conditions and the energy use were validated using on-site measurement. The developed tools are currently only working with a one-way coupling (Mauree et al. 2018) and hence still need to be improved to provide a better representation and local-scale feedback in meteorological models. 5 Linking Neighborhoods into Sustainable Energy Systems 99

5.1.2 Urban Energy Simulation

Simulating the energy demand at urban scale is a challenging task due to the climate uncertainty and the influence of urban microclimate. The methods used to model the energy demand of a city, can be classified into two approaches: top-down and bottom-up. Top-down models often use data such as energy demand, CO2 emissions, financial aspects, etc., at national, regional or urban scale in order to quantify the energy performance of single buildings or building stocks (Kavgic et al. 2010). In contrast, bottom-up models estimate the energy consumption at building scale or even finer (such as a zone in a building) and extend the model to consider a city, a region, or a country (Swan and Ugursal 2009). Several programs exist to quantify the energy performance of edifices, from the building to the city scale, such as for example CitySim (Robinson et al. 2009), Urban Modeling Interface (UMI) (Reinhart et al. 2013) and SIMSTADT (Nouvel et al. 2015). The main problem when working with these tools is that each urban simulation engine has its own tailor-made data model, and it is quite difficult to make the models commu- nicate with each other. A first step to address this issue was the creation of an Urban Energy Information standard such as the Application Domain Extension (ADE) of the CityGML urban information model (Nouvel et al. 2015; Coccolo and Kämpf 2015). The input data for the analyses should be as precise as possible, from the climate to the occupancy profile. They can be categorized into three typologies: (i) Must-have, (ii) Relevant-to-have, and (iii) Nice-to-have. The Must-have (e.g., year of construction, function, refurbishment, and residence type) should be as precise as possible, since if wrongly entered, they cause a major error of up to 30% in the estimation of the energy demand (Nouvel et al. 2017). Similarly, the occupancy profile of the buildings plays a vital role. Normally, the relevant data can only be obtained from a large database of information related to the corresponding building stock, which is not always available. In order to overcome this difficulty, a new direction is the use of an archetype, i.e., a geometrical abstraction of the urban environment. Generally, when working with an existing city, all its physical properties need to be known (climate, geometry, building type, etc.) to perform the analyses. By contrast, when working with archetypes (Ratti et al. 2003), the geometry can be simplified, by modeling a city as a function of the following parameters: (i) Plot Ratio (-), (ii) Site Coverage (%), and (iii) Form Factor (-). By working with archetypes, the computation time required is drastically reduced, which allows a new way to visualize the city’s energy behavior and to generate solutions for sustainable urban design. Optimization of the energy demand of buildings is essential in the planning of a new district and in the choice of renovation scenarios for existing building stocks. The related energy system design process will be discussed in detail in Sects. 5.1.3– 5.1.5. In addition, it is important to consider the influence of building stock on the urban microclimate; especially in the context of outdoor thermal comfort. Urban energy simulation needs to be coupled with energy system design and urban 100 A.T.D. Perera et al.

Fig. 5.3 Optimization, by new hybrid evolutionary algorithm (CMA-ES/HDE), of the EPFL campus in Lausanne, focusing on the energy demand and the outdoor thermal comfort (Coccolo 2017) planning in order to link the building stock in an effective way to the energy infrastructure. A holistic design platform combining different aspects is shown in Fig. 5.3.

5.1.2.1 Sustainable Methods to Cater the Energy Demand

Distributed energy systems incorporating renewable energy technologies will play a vital role in the energy transition. Designing such systems for cities is challenging due to the complexity of urban conﬁgurations that inﬂuence the wind pattern and solar irradiation, especially with regard to building envelope elements. Both integration of renewable energy technologies at the building scale and energy system upgrading by installing energy storage and energy conversion methods such as heat pumps, etc., are essential in this context. This implies assessment of the renewable energy integration potential at the building scale, energy system design at neighbourhood and urban scale, and assessment of the energy systems at the neighbourhood and urban scale.

5.1.2.2 Integration of Renewable Technologies at the Building Scale

More than one-third of buildings in the EU27, Switzerland and Norway were built before 1960, when no energy saving policy was in place (Nolte and Strong 2011). To achieve a Nearly Zero annual energy balance, energy needs must be diminished and matched with distributed renewable energy generation: this will reduce energy distribution dispersions and increase awareness, by transforming energy consumers into energy “prosumers”. Building envelopes, when appropriately equipped, can become active and generate energy: as such, envelope surfaces constitute an important resource for an effective transition to renewables. Thermal energy can be produced from Building-Integrated Solar Collectors (BIST) and electricity from Building-Integrated Photovoltaic modules (BIPV), when there is satisfactory solar 5 Linking Neighborhoods into Sustainable Energy Systems 101 radiation unshaded by the surroundings: the suitable envelope area for solar collection corresponds roughly to 60% of roof areas and 20% of façade areas (IEA 2002). The scientific literature suggests that up to 50% of low and medium temperature heat in Europe could be delivered by solar thermal technology by 2030 (ESTTP 2009). PV modules could cover between 15 and 60% of the electricity demand in IEA countries. More than half of this capacity is expected to be installed on buildings by 2050 (IEA 2014), helped by the continuous reduction in price of PV technologies (Zhang et al. 2014). In addition, a small amount of renewable electricity may be generated by Building-Integrated Wind Turbines (BIWT), which can provide up to 15% of the electricity demand of a single building (Bošnjaković 2013). The renewable energy potential can be assessed at the territorial scale and refined further to the scale of a single building. The solar energy potential is usually calculated from the solar radiation impinging on a given surface multiplied by its conversion efficiency into useful energy (Fath et al. 2015): depending on the technology and on the external air temperature, solar collectors convert up to 80% of solar radiation into thermal energy and PV modules convert up to 22% of solar radiation into electricity (Probst and Roecker 2012). Solar radiation is usually measured in meteorological stations then modeled according to expected shading, tilt from the horizontal plane and azimuthal orientation of the target surface. Wind energy potential, on the other hand, depends on wind velocity as well as on tip-speed ratio and pitch angle as a function of the turbine type (Campos-Arriaga 2009). Wind speed, like solar radiation, is measured in meteorological stations and evaluated through specific fluid dynamics models.

5.1.3 Integrated Energy Systems Based on Renewable Energy Technologies

Integrated energy systems such as energy hubs, smart microgrids, etc., are getting popular as a method to integrate larger fractions of non-dispatchable energy technologies such as solar PV and wind energy (Perera et al. 2015, 2016, 2017a, b; Guen et al. 2018; Kuehner et al. 2017). It is essential to keep a good balance between renewable energy technologies, energy storage and dispatchable energy sources (Perera et al. 2012, 2013a, b). Hence, a number of different optimization methods have been proposed to design distributed energy systems using various methods (Connolly et al. 2010). In addition to design optimization, it is important to locate the energy system considering the distributed energy demands and opportunities for on-site power generation (Max Bittel et al. 2017). Finally, the energy system should be operated considering the variations in demand, renewable energy potential and grid conditions (Timothée et al. 2017). However, so far, the main focus has been on the optimization of the energy system. 102 A.T.D. Perera et al.

Incorporating building-integrated energy technologies into the energy infrastructure and planning energy efficient cities combining both urban planning and energy system optimization are challenging tasks. This is mainly due to insufficient information available for urban energy simulation and the fact that extensive computational resources are required for simultaneous optimization of the building stock and the energy system. The process becomes even more difficult due to the impact of the microclimate on the energy demand and generation. A detailed computational platform combining urban climate, building simulation, and energy system optimization would be helpful as suggested in Mauree et al. (2018), Perera et al. (2018). However, the best solution for optimizing urban configuration and energy system design is likely to be a simplification of both urban simulation and energy system optimization parts. Urban archetypes can be used to represent the complex urban morphologies (Ratti et al. 2003) (as suggested in Sect. 5.1.2) while meta-models can be used to simplify the optimization process as suggested by Perera et al. (2017), Fig. 5.4.

5.1.4 Assessment of Distributed Energy Systems

Urban energy planning is a broad subject which does not end with optimization. According to Manfren et al. (2011), it consists of ﬁve major phases starting with the collection of basic data, their preprocessing, and energy system design. Once the energy system has been designed, it is important to go through a post-processing phase, evaluating aspects related to energy efﬁciency, economy, environmental

Electricity Demand of the Appliances

SPV Panels

Wind Turbines

Dense building stock City center

Internal combusƟon Engine

Less dense building stock Periphery Detailed informaƟon AbstracƟng the Energy system (at the building level) building model designing problem about building stock using architypes

Fig. 5.4 Developing a computational platform for the planning of energy sustainable neighborhoods 5 Linking Neighborhoods into Sustainable Energy Systems 103

Fig. 5.5 Steps to be followed in urban energy planning. The process begins with identifying the energy demand and energy potentials for renewable energy technologies. Subsequently, energy system is designed considering several objective functions. The set of Pareto solutions obtained from the Pareto optimization is ranked based on the priorities and design requirements of the application impact, and social acceptance as suggested in Fig. 5.5. A number of different methods have been proposed to carry out this task, which will ideally be combined with an impact assessment including a life cycle assessment and local DG planning. Developing design tools to optimize energy systems considering urban context is important. However, post-processing and impact assessment phases should not be neglected in order to complete the process. Combining the design and the post-processing phase is always challenging, especially when trying to come up with the ﬁnal system design. Multiple criteria need to be considered simultaneously in the process. This is where multi-criterion decision-making comes into play. Different methods based on Fuzzy TOPSIS, Analytical Hierarchical Process, etc., can be used (Perera et al. 2013, 2017).

5.1.5 Social Acceptance for Renewable Energy Integration

The deployment of building-integrated renewables is conditioned by their social acceptance, resulting from the local urban context (Probst and Roecker 2015) (Fig. 5.6). This usually cannot be controlled through a linear relationship as explained in Fig. 5.6. A fundamental influence on social acceptance is location-specific socio-cultural sensitivity, linked with the historical development of the buildings and their appearance, as well as the cultural and vernacular heritage they represent. Other relevant aspects are the traditional space use made by the citizens, the presence of symbols, icons, landmarks, monuments or institutions in which inhabitants reflect their habits, the common predilection, and diffuse well-being for that place. Sensitivity depends also on the resilience of the built environment to the intrusion of renewable systems and the capacity to accumulate their proliferation without losing value: residential and industrial areas are typically 104 A.T.D. Perera et al.

Fig. 5.6 Example of comprehensive solar development planning, considering social acceptance in relation with its various parameters: Hollande district, Geneva Switzerland. From bottom to top: (i) urban sensitivity issued from land use; (ii) visual interest inferred from a census of remarkable viewpoints or crowd-sourced photographs; (iii) annual solar radiation on a gradient color scale; (iv) visibility of envelope surfaces from the public space; (v) resulting integration strategy: standard modules on non-visible surfaces and high-end products on visible surfaces 5 Linking Neighborhoods into Sustainable Energy Systems 105 less sensitive than historical settlements. The level of coherence of renewable system designs within existing buildings is determined in relation with (i) the system size, position, and layout; (ii) the visible materials, including textures and colors; (iii) the modular pattern of the single elements composing the array; and (iv) the connections and joints between the system components. Such features can be summarized with the designation of overall system quality (Probst and Roecker 2011). Another relevant factor for social acceptance is the visibility of the renewable systems from the public space. This depends on the visual interest of the zone linked with the number of potential observers, the nature and duration of their viewing experience as well as with the scenic attractiveness of the urban landscape: for instance, visual interest can be assessed by observing the distribution of photographs on a territory (Florio et al. 2017). Beyond this social component, visibility depends also on the impact of the renewable systems on the visual field of potential observers, influenced by the local topography, the physical features of the system and its surrounding elements as well as the meteorological, atmospheric, and lighting conditions. It should not be forgotten that building envelopes represent the visible interface of buildings exposed to the public space. Visibility assessment methods of envelope surfaces that could host renewable energy generation plants are needed in the planning or predesign phase, knowing that solar technologies can be coplanar with the relative envelope surface, which is seldom the case for wind turbines. These methods include viewsheds and bi/tridimensional isovists, which identify the surfaces that can eventually be visible from the public space and the number of viewing locations (Hurtado et al. 2004; Fernandez-Jimenez et al. 2015). A finer estimation, providing a magnitude scale of visibility, is constituted by indicators based on solid angles, computing the projection of target surfaces on a spherical field of view (Minelli et al. 2014; Rodrigues et al. 2010). In addition, psy- chophysical considerations on visual acuity can improve assessment methods based on perceptual evidence (Florio et al. 2016). As an alternative, virtual, and aug- mented reality environments allow immersing observers in a simulated renewable project and tracking their interaction on a computer appliance (Lizcano et al. 2017).

5.1.6 Conclusions and Future Perspectives

This book chapter provides the “big picture” on different important aspects that need to be considered when linking buildings with sustainable energy technologies. It highlights a number of different challenges related to determining energy demand and renewable energy potential in urban context and promising paths to overcome them. Combining the energy system design process and urban planning in order to reach toward energy sustainable neighborhoods is a multi-faceted task. A computational platform that combines several modules related to urban climate, urban simulation, and energy system optimization would be beneficial. However, 106 A.T.D. Perera et al. there will always be the difficulty of maintaining the information flow. Finally, this chapter has highlighted the importance of assessment of energy infrastructure beyond the design process of energy efficient neighborhoods or districts. Social acceptance of renewable energy integration plays an important role, which might well be beyond the level of control of urban planners or energy system designers.

References

Alves CA, Duarte DHS, Gonçalves FLT (2016) Residential buildings’ thermal performance and comfort for the elderly under climate changes context in the city of São Paulo Brazil. Energy Build 114:62–71. https://doi.org/10.1016/j.enbuild.2015.06.044 Ashie Y, Ca VT, Asaeda T (1999) Building canopy model for the analysis of urban climate. J Wind Eng Ind Aerodyn 81:237–248. https://doi.org/10.1016/S0167-6105(99)00020-3 Barbosa R, Vicente R, Santos R (2015) Climate change and thermal comfort in Southern Europe housing: a case study from Lisbon. Build Env 92:440–451. https://doi.org/10.1016/j.buildenv. 2015.05.019 Belcher S, Hacker J, Powell D (2005) Constructing design weather data for future climates. Build Serv Eng Res Technol 26:49–61. https://doi.org/10.1191/0143624405bt112oa Bošnjaković M (2013) Wind power buildings integration. J Mech Eng Autom 3. https://doi.org/10. 17265/2159-5275/2013.04.005 Bueno B, Hidalgo J, Pigeon G, Norford L, Masson V (2013) Calculation of air temperatures above the urban canopy layer from measurements at a rural operational weather station. J Appl Meteorol Climatol 52:472–483. https://doi.org/10.1175/JAMC-D-12-083.1 Campos-Arriaga L (2009) Wind energy in the built environment : a design analysis using CFD and wind tunnel modelling approach. The University of Nottingham Chow DHC, Li Z, Darkwa J (2013) The effectiveness of retroﬁtting existing public buildings in face of future climate change in the hot summer cold winter region of China. Energy Build 57:176–186. https://doi.org/10.1016/j.enbuild.2012.11.012 Christensen J, Kjellström E, Giorgi F, Lenderink G, Rummukainen M (2010) Weight assignment in regional climate models. Clim Res 44:179–194 Coccolo S (2017) Bioclimatic design of sustainable campuses using advanced optimisation methods Coccolo S, Kämpf J (2015) Urban energy simulation based on a new data model paradigm: the CityGML application domain extension energy. A case study in the EPFL campus of Lausanne. In: 14th International conference IBPSA—building simulation 2015 BS 2015 conference proceedings Connolly D, Lund H, Mathiesen BV, Leahy M (2010) A review of computer tools for analysing the integration of renewable energy into various energy systems. Appl Energy 87:1059–1082. https://doi.org/10.1016/j.apenergy.2009.09.026 CORDEX n.d. http://cordex.org/. Accessed 27 Feb 2016 European Solar Thermal Technology Platform (ESTTP) (2009) Solar heating and cooling for a sustainable energy future in Europe Fant C, Adam Schlosser C, Strzepek K (2016) The impact of climate change on wind and solar resources in Southern Africa. Appl Energy 161:556–564. https://doi.org/10.1016/j.apenergy. 2015.03.042 Fath K, Stengel J, Sprenger W, Wilson HR, Schultmann F, Kuhn TE (2015) A method for predicting the economic potential of (building-integrated) photovoltaics in urban areas based on hourly radiance simulations. Sol Energy 116:357–370. https://doi.org/10.1016/j.solener. 2015.03.023 5 Linking Neighborhoods into Sustainable Energy Systems 107

Fernandez-Jimenez LA, Mendoza-Villena M, Zorzano-Santamaria P, Garcia-Garrido E, Lara-Santillan P, Zorzano-Alba E et al (2015) Site selection for new PV power plants based on their observability. Renew Energy 78:7–15. https://doi.org/10.1016/j.renene.2014.12.063 Field CB, Barros V, Stocker TF, Dokken DJ, Ebi KL, Mastrandrea M et al (2012) Summary for policymakers. In: Managing the risks of extreme events and disasters to advance climate change adaptation. Cambridge University Press, Cambridge, UK, New York, NY, USA Fisk WJ (2015) Review of some effects of climate change on indoor environmental quality and health and associated no-regrets mitigation measures. Build Env 86:70–80. https://doi.org/10. 1016/j.buildenv.2014.12.024 Florio P, Roecker C, Munari Probst MC, Scartezzini J-L (2016) Visibility of building exposed surfaces for the potential application of solar panels: a photometric model. In: Biljecki F, Tourre V (eds) Eurographics workshop urban data model. Vis., Liège. https://doi.org/10.2312/ udmv.20161419 Florio P, Probst MCM, Schüler A, Scartezzini J-L (2017) Visual prominence vs architectural sensitivity of solar applications in existing urban areas: an experience with web-shared photos. Energy Procedia 122. https://doi.org/10.1016/j.egypro.2017.07.437 Fowler HJ, Blenkinsop S, Tebaldi C (2007) Linking climate change modelling to impacts studies: recent advances in downscaling techniques for hydrological modelling. Int J Climatol 27:1547–1578. https://doi.org/10.1002/joc.1556 Garuma GF (2017) Review of urban surface parameterizations for numerical climate models. Urban Clim 2017. https://doi.org/10.1016/j.uclim.2017.10.006 Giorgetta MA, Jungclaus J, Reick CH, Legutke S, Bader J, Böttinger M et al (2013) Climate and carbon cycle changes from 1850 to 2100 in MPI-ESM simulations for the coupled model intercomparison project phase 5. J Adv Model Earth Syst 5:572–597. https://doi.org/10.1002/ jame.20038 Giorgi F (2006) Regional climate modeling: status and perspectives. J Phys IV Proc 139:18. https://doi.org/10.1051/jp4:2006139008 Guen ML, Mosca L, Perera ATD, Coccolo S, Mohajeri N, Scartezzini J-L (2018) Improving the energy sustainability of a Swiss village through building renovation and renewable energy integration. Energy Build 158:906–923. https://doi.org/10.1016/j.enbuild.2017.10.057 Hurtado JP, Fernández J, Parrondo JL, Blanco E (2004) Spanish method of visual impact evaluation in wind farms. Renew Sustain Energy Rev 8:483–491. https://doi.org/10.1016/j. rser.2003.12.009 ICLEI U. UN-Habitat (2009) Sustainable Urban Energy Planning. https://unhabitat.org/books/ sustainable-urban-energy-planning/. Accessed 15 Oct 2018 International Energy Agency IEA (2002) Potential for building integrated photovoltaics International Energy Agency IEA (2014) Technology roadmap: solar photovoltaic energy IPCC (2007) Climate change 2007: synthesis report. Contribution of working groups I, II and III to the fourth assessment report of the intergovernmental panel on climate change. IPCC, Geneva, Switzerland Kalvelage K, Passe U, Rabideau S, Takle ES (2014) Changing climate: the effects on energy demand and human comfort. Energy Build 76:373–380. https://doi.org/10.1016/j.enbuild. 2014.03.009 Kao S-C, Sale MJ, Ashfaq M, Uria Martinez R, Kaiser DP, Wei Y et al (2015) Projecting changes in annual hydropower generation using regional runoff data: an assessment of the United States federal hydropower plants. Energy 80:239–250. https://doi.org/10.1016/j.energy.2014.11.066 Karimpour M, Belusko M, Xing K, Boland J, Bruno F (2015) Impact of climate change on the design of energy efﬁcient residential building envelopes. Energy Build 87:142–154. https://doi. org/10.1016/j.enbuild.2014.10.064 Kavgic M, Mavrogianni A, Mumovic D, Summerﬁeld A, Stevanovic Z, Djurovic-Petrovic M (2010) A review of bottom-up building stock models for energy consumption in the residential sector. Build Env 45:1683–1697. https://doi.org/10.1016/j.buildenv.2010.01.021 Kershaw T, Eames M, Coley D (2011) Assessing the risk of climate change for buildings: a comparison between multi-year and probabilistic reference year simulations. Build Env 46:1303–1308. https://doi.org/10.1016/j.buildenv.2010.12.018 108 A.T.D. Perera et al.

Kjellström E, Nikulin G, Hansson U, Strandberg G, Ullerstig A (2011) 21st century changes in the European climate: uncertainties derived from an ensemble of regional climate model simulations. Tellus A 63:24–40 Kohler M, Blond N, Clappier A (2016) A city scale degree-day method to assess building space heating energy demands in Strasbourg Eurometropolis (France). Appl Energy 184:40–54. https://doi.org/10.1016/j.apenergy.2016.09.075 Krpo A, Salamanca F, Martilli A, Clappier A (2010) On the impact of anthropogenic heat ﬂuxes on the urban boundary layer: a two-dimensional numerical study. Bound-Layer Meteorol 136:105–127. https://doi.org/10.1007/s10546-010-9491-2 Kuehner AL, Mdeihli N, Coccolo S, Perera ATD, Mohajeri N, Scartezzini J-L (2017) Extending building integrated photovoltaics (BiPV) using distributed energy hubs. A case study in Cartigny, Switzerland. Energy Procedia 122:487–492. https://doi.org/10.1016/j.egypro.2017. 07.299 Lizcano PE, Manchado C, Gomez-Jauregui V, Otero C (2017) Virtual reality to assess visual impact in wind energy projects. In: Eynard B, Nigrelli V, Oliveri S, Peris-Fajarnes G, Rizzuti S (eds) Advances on mechanics, design engineering and manufacturing. Springer, Cham, pp 717–725. https://doi.org/10.1007/978-3-319-45781-9_72 Manfren M, Caputo P, Costa G (2011) Paradigm shift in urban energy systems through distributed generation: methods and models. Appl Energy 88:1032–1048. https://doi.org/10.1016/j. apenergy.2010.10.018 Martilli A (2007) Current research and future challenges in urban mesoscale modelling. Int J Climatol 27:1909–1918. https://doi.org/10.1002/joc.1620 Mauree D, Nadège B, Clappier A, Kämpf JH, Scartezzini J-L (2015) Evaluation of building energy use: from the urban to the building scale. In: Proceedings of 9th international conference on urban climate, Toulouse Mauree D, Coccolo S, Kaempf J, Scartezzini J-L (2017) Multi-scale modelling to evaluate building energy consumption at the neighbourhood scale. PLoS One 12:e0183437. https://doi. org/10.1371/journal.pone.0183437 Mauree D, Coccolo S, Perera ATD, Nik V, Scartezzini J-L, Naboni E (2018a) A new framework to evaluate urban design using urban microclimatic modeling in future climatic conditions. Sustainability 10:1134. https://doi.org/10.3390/su10041134 Mauree D, Blond N, Clappier A (2018) Multi-scale modeling of the urban meteorology: integration of a new canopy model in the WRF model. EarthArXiv 2018. https://doi.org/10. 17605/osf.io/w89cj Max Bittel H, Perera ATD, Mauree D, Scartezzini J-L (2017) Locating multi energy systems for a neighborhood in Geneva using K-means CLUSTERING. Energy Procedia 122:169–174. https://doi.org/10.1016/j.egypro.2017.07.341 Meehl GA, Stocker TF, Collins W, Friedlingstein P, Gaye A, Gregory J et al (2007) Global climate projections climate change 2007: the physical science basis. In: Qin D, Manning M, Chen Z, Marquis M, Averyt K, Tignor M, Mill HL (eds) Contribution of working group I to the fourth assessment report of the intergovernmental panel on climate change. Cambridge University Press, New York, pp 747–845 Minelli A, Marchesini I, Taylor FE, De Rosa P, Casagrande L, Cenci M (2014) An open source GIS tool to quantify the visual impact of wind turbines and photovoltaic panels. Env Impact Assess Rev 49:70–78. https://doi.org/10.1016/j.eiar.2014.07.002 Munari Probst MC, Roecker C (2011) Urban acceptability of building integrated solar systems: LESO-QSV approach. Eurosun 2011, ISES 2011 Munari Probst MC, Roecker C (eds) (2012) Solar energy systems in architecture. DA 2. IEA SHC Task 41 Munari Probst MC, Roecker C (2015) Solar energy promotion & urban context protection: LESO-QSV (Quality—Site—Visibility) method. In: 31th International PLEA conference Nik VM (2010) Climate simulation of an attic using future weather data sets—statistical methods for data processing and analysis. Licentiate Thesis. Chalmers University of Technology 5 Linking Neighborhoods into Sustainable Energy Systems 109

Nik VM, Sasic Kalagasidis A, Kjellström E (2012a) Assessment of hygrothermal performance and mould growth risk in ventilated attics in respect to possible climate changes in Sweden. Build Env 55:96–109. https://doi.org/10.1016/j.buildenv.2012.01.024 Nik VM, Sasic Kalagasidis A, Kjellström E (2012b) Statistical methods for assessing and analysing the building performance in respect to the future climate. Build Env 53:107–118. https://doi.org/10.1016/j.buildenv.2012.01.015 Nolte I, Strong D (2011) Europe’s buildings under the microscope Nouvel R, Brassel K-H, Bruse M, Duminil E, Coors V, Eicker U et al (2015) SIMSTADT, a new worflow-driven urban energy simulation platform for CityGML City Models. In: CISBAT international conference, pp 889–894 Nouvel R, Zirak M, Coors V, Eicker U (2017) The influence of data quality on urban heating demand modeling using 3D city models. Comput Env Urban Syst 64:68–80. https://doi.org/10. 1016/j.compenvurbsys.2016.12.005 Oke TR (1967) City size and the urban heat island. Atmos Env 1973 7:769–779. https://doi.org/10. 1016/0004-6981(73)90140-6 Oke TR (1982) The energetic basis of the urban heat island. Q J R Meteorol Soc 108:1–24. https:// doi.org/10.1002/qj.49710845502 Pakkala TA, Köliö A, Lahdensivu J, Kiviste M (2014) Durability demands related to frost attack for Finnish concrete buildings in changing climate. Build Env 82:27–41. https://doi.org/10. 1016/j.buildenv.2014.07.028 Perera ATD, Wickremasinghe DMIJ, Mahindarathna DVS, Attalage RA, Perera KKCK, Bartholameuz EM (2012) Sensitivity of internal combustion generator capacity in standalone hybrid energy systems. Energy 39:403–411. https://doi.org/10.1016/j.energy.2011.12.039 Perera ATD, Attalage RA, Perera KKCK, Dassanayake VPC (2013a) Converting existing internal combustion generator (ICG) systems into HESs in standalone applications. Energy Convers Manag 74:237–248. https://doi.org/10.1016/j.enconman.2013.05.022 Perera ATD, Attalage RA, Perera KKCK, Dassanayake VPC (2013b) Designing standalone hybrid energy systems minimizing initial investment, life cycle cost and pollutant emission. Energy 54:220–230. https://doi.org/10.1016/j.energy.2013.03.028 Perera ATD, Attalage RA, Perera KKCK, Dassanayake VPC (2013c) A hybrid tool to combine multi-objective optimization and multi-criterion decision making in designing standalone hybrid energy systems. Appl Energy 107:412–425. https://doi.org/10.1016/j.apenergy.2013.02. 049 Perera ATD, Madusanka AN, Attalage RA, Perera KKCK (2015) A multi criterion analysis for renewable energy integration process of a standalone hybrid energy system with internal combustion generator. J Renew Sustain Energy 7:043128. https://doi.org/10.1063/1.4928684 Perera ATD, Mauree D, Scartezzini JL, Nik VM (2016) Optimum design and control of grid integrated electrical hubs considering lifecycle cost and emission. In: 2016 IEEE international energy conference ENERGYCON, pp 1–6. https://doi.org/10.1109/energycon.2016.7513968 Perera ATD, Nik VM, Mauree D, Scartezzini J-L (2017a) Electrical hubs: an effective way to integrate non-dispatchable renewable energy sources with minimum impact to the grid. Appl Energy 190:232–248. https://doi.org/10.1016/j.apenergy.2016.12.127 Perera ATD, Mauree D, Scartezzini J-L (2017b) The energy hub concept applied to a case study of mixed residential and administrative buildings in Switzerland. Energy Procedia 122:181–186. https://doi.org/10.1016/j.egypro.2017.07.342 Perera ATD, Wickramasinghe U, Nik VM, Scartezzini J-L (2017c) Optimum design of distributed energy hubs using hybrid surrogate models (HSM). Energy Procedia 122:187–192. https://doi. org/10.1016/j.egypro.2017.07.343 Perera ATD, Nik VM, Mauree D, Scartezzini J-L (2017d) An integrated approach to design site specific distributed electrical hubs combining optimization, multi-criterion assessment and decision making. Energy 134:103–120. https://doi.org/10.1016/j.energy.2017.06.002 Perera ATD, Coccolo S, Scartezzini J-L, Mauree D (2018) Quantifying the impact of urban climate by extending the boundaries of urban energy system modeling. Appl Energy 222:847– 860. https://doi.org/10.1016/j.apenergy.2018.04.004 110 A.T.D. Perera et al.

Pigeon G, Zibouche K, Bueno B, Le Bras J, Masson V (2014) Improving the capabilities of the town energy balance model with up-to-date building energy simulation algorithms: an application to a set of representative buildings in Paris. Energy Build 76:1–14. https://doi.org/ 10.1016/j.enbuild.2013.10.038 Prudhomme C, Wilby RL, Crooks S, Kay AL, Reynard NS (2010) Scenario-neutral approach to climate change impact studies: application to flood risk. J Hydrol 390:198–209. https://doi.org/ 10.1016/j.jhydrol.2010.06.043 Pryor SC, Schoof JT, Barthelmie RJ (2006) Winds of change?: projections of near-surface winds under climate change scenarios. Geophys Res Lett 33:L11702. https://doi.org/10.1029/ 2006GL026000 Ratti C, Raydan D, Steemers K (2003) Building form and environmental performance: archetypes, analysis and an arid climate. Energy Build 35:49–59. https://doi.org/10.1016/S0378-7788(02) 00079-8 Reinhart CF, Davila CC (2016) Urban building energy modeling—a review of a nascent field. Build Env 97:196–202. https://doi.org/10.1016/j.buildenv.2015.12.001 Reinhart C, Dogan T, Jakubiec A, Rakha T, Sang A (2013) UMI—an urban simulation environment for building energy use, daylighting and walkability. In: Reinhart CF, Dogan T, Jakubiec JA, Rakha T, Sang A (eds) 13th Conference on International Buildings Performance Simulation Association Chambéry France, Massachusetts Institute of Technology Department of Architecture, August 26–28, pp 476–483 Remund J, Müller S, Kunz S. (2015) Meteonorm. Global Metereological Database. Version 7 Robinson D, Haldi F, Kämpf J, Leroux P (2009) CitySim: comprehensive micro-simulation of resource flows for sustainable urban planning. In: Proceedings of eleventh international IBPSA conference, Glasgow Rodrigues M, Montañés C, Fueyo N (2010) A method for the assessment of the visual impact caused by the large-scale deployment of renewable-energy facilities. Env Impact Assess Rev 30:240–246. https://doi.org/10.1016/j.eiar.2009.10.004 Salamanca F, Martilli A (2010) A new building energy model coupled with an urban canopy parameterization for urban climate simulations—part II. Validation with one dimension off-line simulations. Theory Appl Climatol 99:345–356. https://doi.org/10.1007/s00704-009-0143-8 Samuelsson P, Jones CG, Willen U, Ullerstig A, Gollvik S, Hansson U et al (2011) The Rossby Centre regional climate model RCA3: model description and performance. Tellus A 63:4–23 Samuelsson P, Gollvik S, Jansson C, Kupiainen M, Kourzeneva E, de Berg WJ van (2015) The surface processes of the Rossby Centre regional atmospheric climate model (RCA4). Swedish Meteorological and Hydrological Institute (SMHI) Seljom P, Rosenberg E, Fidje A, Haugen JE, Meir M, Rekstad J et al (2011) Modelling the effects of climate change on the energy system—a case study of Norway. Energy Policy 39:7310– 7321. https://doi.org/10.1016/j.enpol.2011.08.054 Shibuya T, Croxford B (2016) The effect of climate change on office building energy consumption in Japan. Energy Build 117:149–159. https://doi.org/10.1016/j.enbuild.2016.02.023 Swan LG, Ugursal VI (2009) Modeling of end-use energy consumption in the residential sector : a review of modeling techniques 13:1819–1835. https://doi.org/10.1016/j.rser.2008.09.033 The Global Risks Report (2016) Geneva, Switzerland: World Economic Forum Timothée C, Perera ATD, Scartezzini J-L, Mauree D (2017) Optimum dispatch of a multi-storage and multi-energy hub with demand response and restricted grid interactions. Energy Procedia 142:2864–2869. https://doi.org/10.1016/j.egypro.2017.12.434 Wang H, Chen Q (2014) Impact of climate change heating and cooling energy use in buildings in the United States. Energy Build 82:428–436. https://doi.org/10.1016/j.enbuild.2014.07.034 de Wilde P, Coley D (2012) The implications of a changing climate for buildings. Build Env 55:1– 7. https://doi.org/10.1016/j.buildenv.2012.03.014 Zhang HL, Van Gerven T, Baeyens J, Degrève J, Ve J (2014) Photovoltaics: reviewing the European feed-in-tariffs and changing PV efficiencies and costs. Sci World J 2014:404913. https://doi.org/10.1155/2014/404913 Chapter 6 Future Weather Data for Dynamic Building Energy Simulations: Overview of Available Data and Presentation of Newly Derived Data for Belgium

Delphine Ramon , Karen Allacker , Nicole P. M. van Lipzig, Frank De Troyer and Hendrik Wouters

Abstract As buildings have a relatively long life span, it is important to consider climate change in energy performance modelling. Good quality weather data are needed to obtain accurate results. This chapter discusses widely used methods to predict future weather data (dynamical downscaling, stochastic weather generators and morphing) and provides an overview of available weather datasets (multi-year, typical years, extreme years and representative years) for building simulations. A Flemish office building is used for a comparative analysis of the estimated heating and cooling load making use of 1-year weather files (typical and extreme future climate conditions) derived from a recently developed convection-permitting climate model for Belgium. Climate models and weather generators are identified as the most preferred for the estimation of the average energy consumption and thermal comfort in average and extreme situations. Climate models have the advantage to better represent extreme weather events and climate differences due to territorial settings, while weather generators can generate multiple climate realizations. A combination of a typical year with an extreme cold and extreme warm year was found to result in an overall good representation of the energy need for heating and cooling in average and extreme weather conditions. Further, the influence of the methodological choices to extract 1-year weather files (typical or extreme years) from the 30-year climate data is highlighted as different results were

D. Ramon (&) Á K. Allacker Á F. De Troyer Faculty of Engineering Science, Department of Architecture, KU Leuven, Louvain, Belgium e-mail: [email protected] N. P. M. van Lipzig Faculty of Science, Department of Earth and Environmental Sciences, KU Leuven, Louvain, Belgium H. Wouters Faculty of Bioscience Engineering, Department of Forest and Water Management, UGent, Ghent, Belgium

© Springer Nature Singapore Pte Ltd. 2019 111 E. Motoasca et al. (eds.), Energy Sustainability in Built and Urban Environments, Energy, Environment, and Sustainability, https://doi.org/10.1007/978-981-13-3284-5_6 112 D. Ramon et al. obtained when different meteorological variables were considered for the creation of the 1-year ﬁles.

Keywords Future weather ﬁles Á Climate modelling Á Built environment Convection-permitting climate model Á Typical year Á Extreme year

6.1 Introduction

People spend a lot of time in buildings protecting them from the outdoor weather. These buildings are affected to an important extent by the local weather. The heatwave of 2003 caused thousands of excess deaths in France only (Vandentorren et al. 2006), and indoor thermal discomfort is considered as one factor in this excess of deaths. Towards the future, an increase in temperature is expected and in particular an increase in frequency of heatwave periods (Wouters et al. 2017; Kovats et al. 2014; Berger et al. 2014). Heatwaves are also becoming more severe in urban areas due to the urban heat island effect (Wouters et al. 2017), likely leading to comfort and well-being problems for occupants at a regular base and/or to an increase in electricity consumption for active cooling (Crawley 2008). In the past decade, various researchers worldwide investigated the impact of climate change on the building energy performance (Berger et al. 2014; Andrić et al. 2017; Shen 2017; Shibuya and Croxford 2016; Kershaw et al. 2011; Brotas and Nicol 2017; Chow and Levermore 2010; de Wilde and Tian 2010; Farrou et al. 2014; Holmes and Hacker 2007; MBE KTN 2013; Nik and Sasic Kalagasidis 2013;RICS2015). Buildings should be resilient to these climate changes without leading to comfort issues or structural damage. To investigate the performance of a building in a context of climate change, the full life cycle of the building should be considered. In order to predict the building energy performance in future, two crucial aspects are needed: (1) an appropriate energy simulation model that can accurately predict building performance and (2) good quality future weather data. Three types of building energy simulations are currently used: static, semi-dynamic and dynamic simulations. In a dynamic energy simulation, a thermal balance is calculated for each time step of the simulation. Based on this balance, the heating and cooling demand as well as the thermal comfort can be defined. The thermal balance takes into account internal heat gains (people, appliances, lighting, etc.), outside weather conditions and the building characteristics (thermal capacity, insulation level, heating and cooling systems, etc.). The interaction of these parameters and their hourly variations are hence considered. A static energy simulation does not take into account all these dependencies or time variations. For example, to calculate the monthly heating demand in static simulations, monthly average temperatures are used. Static energy simulations hence allow to assess the average energy performance of a building (in a rather rough way), but do not allow to determine the thermal comfort as the latter requires hourly data and as the variance of the different elements (e.g. internal gains and solar gains) of the thermal balance 6 Future Weather Data for Dynamic Building Energy Simulations … 113 gain importance. Similarly, this variation is important to investigate a building in extreme weather conditions, e.g. when investigating peak energy demands or thermal discomfort. Semi-dynamic simulations take a shorter time step, e.g. instead of monthly average temperature and solar gains, hourly temperature, hourly direct and indirect solar gains for an average day of each month. For these reasons, dynamic building simulations are more appropriate to assess building resilience towards climate change. Dynamic energy simulations require weather data with at least hourly values, using, for example, an epw format (EnergyPlus 2016). These weather files typically contain information about temperature, radiation (direct and diffuse), wind (direction and speed), rain, snow, humidity and pressure. In addition, it includes among others location-related information such as time zone, elevation or average ground temperatures (EnergyPlus 2016). Depending on the goal of the building energy simulation, weather data should contain information about different weather conditions. If the simulation aims at investigating the overall energy consumption or thermal comfort, weather data representing the typical weather conditions is needed (Barnaby and Crawley 2011; ASHRAE 2017). However, if the aim is to assess the energy consumption or thermal comfort in extreme conditions (i.e. extremely hot or cold), weather data representing those extreme conditions are required (Barnaby and Crawley 2011; ASHRAE 2017). A combination of typical weather data and extreme weather data in one representative dataset could combine both assessments in one simulation (Nik 2016). To size HVAC systems design days or short periods are used (Barnaby and Crawley 2011; ASHRAE 2017). When aiming at climate resilient buildings, future weather data are needed. In order to take into account the inherent uncertainties of climate change, it is important to consider the various possible climate realizations (IPCC 2013). In the past years, appropriate weather data for the assessment of buildings under climate change have been researched (Jentsch et al. 2008, 2013; Jones et al. 2010; Levermoreet al. 2014; Nik 2017; Eames et al. 2011; Struck et al. 2009; Chan 2011; Liu et al. 2016; Narowski et al. 2013). Several future weather files/datasets and methodologies have been developed. The first methodological option is to use a weather file from a different location, more specifically from a location which currently has similar climate conditions as expected for the building location in the future (analogue scenario method) (Belcher et al. 2005; CIBSE 2009). The second option, often used in practice, is to select a year from the past that is warmer than normal to represent the future climate. Third, general circulation models (150– 600 km spatial resolution (IPCC 2013; Jacob et al. 2014)) taking into account future climate scenarios can be used. To be useful for building simulations, downscaling to a relevant spatial resolution for building simulations is needed, e.g. by dynamical downscaling, interpolation, stochastic weather generators or morphing (Belcher et al. 2005; Wilby and Wigley 1997). This paper aims at providing insight in the various methods to predict future weather data and in the available weather datasets appropriate for the Belgian context. The aim is moreover to present recently developed data based on dynamical 114 D. Ramon et al. downscaling of regional climate models for Belgium and to show their potential to be used in building simulations through a comparative analysis with other datasets (e.g. representative) extracted from these models. The chapter is structured as follows. In Sect. 6.2, widely used methods (Sect. 6.2.1) and available datasets (Sect. 6.2.2) are presented, and their appropriateness to predict the building energy performance in a context of climate change (Sect. 6.2.3) is discussed. In the latter section, the appropriateness is linked to the specific simulation goal (heating and cooling demand, thermal comfort evaluation) excluding building simulation for sizing of HVAC. Section 6.3 presents the newly developed future weather data and compares these with other weather datasets extracted from these data. The weather data are moreover applied in the energy simulation of a simple case study (office building). The final section discusses the outlook for weather data and their requirements to assess buildings in the context of climate change.

6.2 Future Weather Data Methods and Data Formats

This section provides an overview of currently available future weather data types. A three-step approach is followed. First, an overview is given of methodologies to obtain future weather data. Second, various weather datasets for building energy simulations are discussed. Finally, the advantages and limitations of the various methods and formats, linked to a certain energy simulation goal, are summarized.

6.2.1 Future Weather Data Methodologies

As discussed before, the analogue scenario method and the selection of a warmer year from the past can be used for building simulations in a future climate context. However, the important role of solar radiation makes it hard to find suited weather files as this depends on the latitude of the location. Moreover, the change of solar radiation seems rather small in current climate change scenarios. Further, both methods assume similar weather behaviour in future as to date while it is expected that frequency of events (e.g. heatwaves) will change in time (Wouters et al. 2017; Kovats et al. 2014; Berger et al. 2014). Therefore, these are not further discussed in this chapter. Ongoing research about future weather data for building simulations typically use of General Circulation Models (GCMs) to gain insights into future climate behaviour (Jentsch et al. 2008, 2013; Jones et al. 2010; Levermore et al. 2014; Nik 2017; Eames et al. 2011; Struck et al. 2009; Chan 2011; Liu et al. 2016). This type of climate models simulates the state and evolution of the atmosphere, including the atmospheric circulation and energy exchanges in terms of radiation, heat and 6 Future Weather Data for Dynamic Building Energy Simulations … 115 moisture. They also simulate the processes related to cloud formation and precipitation, and take into account the interaction with the ocean and the land (IPCC 2013). Climate models are typically used to construct possible trajectories of the future climate. These trajectories particularly take into account the radiative forcing established by the emission of greenhouse gas and aerosols from human activities and natural processes. The most well-known emission scenarios are the Representative Concentration Pathways (RCPs) used by the Intergovernmental Panel on Climate Change (IPCC). They comprise four scenarios, namely, RCP2.6, RCP4.5, RCP6.0 and RCP8.5, which are named after radiative forcing established by the greenhouse gas emissions in the year 2100 relative to pre-industrial values (+2.6, +4.5, +6.0, and +8.5 W/m2, respectively) (Moss et al. 2008). Climate models are also used for studying the climate of the past, e.g. to provide climate information for regions where observations are not available. It should be noted that climate models are not designed to reproduce the order of consecutive weather conditions, as climate models are not constrained by observations. However, they are capable of representing (multi-)decadal climate statistics, such as the mean or variance of temperature of a particular season, or the frequency or mean intensity of heatwaves. Climate models are available for different time spans (up to several decades) but are usually treated for periods of 30 years as the World Meteorological Organization considers that climate statistics converge over this time span (Brisson et al. 2014). GCMs provide climate information on the global scale with a typical spatial resolution of 150–600 km (IPCC 2013; Jacob et al. 2014). When these models are hence used for building energy simulations, the propagation of climate change and related weather extremes at the local level are not taken into account. This is problematic as many buildings are located in cities and are hence affected by urban heat islands. If to be used for building energy simulations, downscaling of the GCMs is needed. To obtain downscaled weather data from GCMs, Belcher et al. (2005) describe four methods: (1) dynamical downscaling; (2) stochastic weather generators; (3) morphing; and (4) interpolation. Wilby and Wigley (1997) further mention regression and weather pattern methods as additional downscaling techniques. In literature review of ongoing research (Jentsch et al. 2008, 2013; Jones et al. 2010; Levermore et al. 2014; Nik 2017; Eames et al. 2011; Struck et al. 2009; Chan 2011; Liu et al. 2016), the first three methods are commonly used to create weather data used in building simulations. These are further discussed in the subsequent paragraphs.

6.2.1.1 Dynamical Downscaling

During the past two decades, GCMs were downscaled to regional climate models (RCMs) by making use of a nesting strategy to obtain climate information at a resolution of 10–100 km. The RCM domain (which covers a smaller region of the 116 D. Ramon et al. world) is nested in the GCM domain. The RCM domain with a ﬁner grid resolution is run with the GCM as initial condition. During the model run, the GCM provides the boundary of the domain climate information. Downscaling of GCMs typically results in an ensemble of RCMs where for each model simulation the same radiative forcing scenarios are used (IPCC 2013). This approach, however, results in inter- nally generated climate variability different for each member due to small pertur- bations of the initial conditions, and hence in different climate realizations in one ensemble. RCMs better represent the regional effects from the orography and the hetero- geneity of the soil, vegetation cover and coastal effects. Some of these RCMs were implemented in building simulations by among others Nik et al. (2016, 2017) and Kikumoto et al. (2015). Nevertheless, RCMs do not yet fully resolve urban effects, while deep atmospheric convection needs to be parameterized as well. The latter two shortcomings of RCMs have been recently resolved by further downscaling the RCMs by means of convection-permitting models (CPMs) providing the mesoscale climate information at a resolution of 1–4 km. As such, cities and deep convection become explicitly resolved (Brisson et al. 2016; Wouters et al. 2016; Kendon et al. 2017), and as such, these CPMs allow to take such information into account. CPMs are hence promising for building simulations as they provide a better representation of local climate (change) compared to RCMs (Prein et al. 2015). CPMs require a high computational time and they are typically available for only one or a few realizations of the future climate. As such, CPMs lack information on the uncertainty of the climate change signal stemming from either physical parameterizations or climate variability. They furthermore are only available for a few regions in the world (e.g. Belgium, UK, Alps, Kilimanjaro region, northwestern Paciﬁc Ocean, Sahel regions (Prein et al. 2015)).

6.2.1.2 Stochastic Weather Generators

Stochastic weather generators are defined by Wilks and Wilby as ‘statistical models which can fill in missing data or produce indefinitely long synthetic weather series by simulating key properties of observed meteorological records (i.e. daily means, variances and covariances, frequencies, extremes, etc.)’ (Wilks and Wilby 1999: p. 329). Originally, weather generators were mainly used in the field of agriculture (e.g. crop production), climate change studies, hydrology and ecology. Later, several weather generators were developed to generate future weather data to be used in building energy simulations, among others by van Paassen and Luo (2002), RUNEOLE by Adelard et al. (2012) and the UKCP09 weather generator by Eames et al. (2010, 2011). Precipitation is most often the primary variable for the stochastic model of the weather generator (Eames et al. 2011; Herrera 2017). Based on the methodology of Hutchinson (Hutchinson 1987) a two-step approach is applied. In the first step, the daily precipitation is modelled based on the current climate data. Based on the fact 6 Future Weather Data for Dynamic Building Energy Simulations … 117 if a day is wet or dry and the weather conditions of the previous day, the other weather variables are generated (mostly temperature and radiation). Parameters such as wind are derived from the latter. Different model parameters for each month are used to include the seasonal variations of the meteorological parameters (Eames et al. 2011; Herrera 2017; Hutchinson 1987). An important advantage of the weather generator methodology is that it allows to integrate the distribution used for the climate change signal and to account for potential changes in weather patterns and climate variability. Some weather generators (e.g. UKCP09 weather generator) furthermore allow to investigate the climate change uncertainty by considering various possible climate realizations (Nik 2017; Eames et al. 2011). Multiple climate realizations can be created by randomly selecting a climate change signal at the beginning of the weather generation (Eames et al. 2011). However, statistical relations between variables as well as distributions are based on baseline data given to the model to generate future data from. Hence, large amounts of data are needed to train the model (Belcher et al. 2005). Further, using stochastic weather generators becomes difficult when statistical relations between meteorological variables are missing in the baseline data because weather events did not happen in that period. The mathematical basis can moreover result in a limited representation and parametrization of climate physics, hence possibly leading to meteorological inconsistencies (Belcher et al. 2005). The weather generator allows a high spatial resolution (e.g. up to 5 km in the case of the UKCP09 weather generator). The climate change signal as such however often has a lower spatial resolution (e.g. 25 km) (Jones et al. 2010). Hence, a spatial variance in the climate change signal in the area of 25 by 25 km caused by territorial settings (e.g. urban versus rural area) is not considered when further downscaling (Wouters et al. 2006, 2017; Berger et al. 2014). With this method, there is moreover no correlation between two time series generated for two adjacent grid cells as a point-based process is used (Jones et al. 2010).

6.2.1.3 Morphing

Belcher et al. (2005) present a methodology for the adjustment of time series towards the future, called ‘morphing’. Current weather data are used as baseline, and monthly climate change signals given by a GCM or RCM are used for morphing the current data. Depending on the climate variable and expression of the climate change signal (absolute, relative), three operations are used to morph data: (1) shifting, (2) scaling and (3) shifting and scaling combined (Belcher et al. 2005). Shifting is applied when the change is expressed absolutely, while scaling is used when a change is relatively expressed. Further, scaling is also used for variables that can be ‘switched off’ (e.g. irradiation). If both the mean and the variance of a variable change over time, shifting and scaling are combined. Shifting changes the mean, while scaling has an inﬂuence on the variance of the weather variable. 118 D. Ramon et al.

Some requirements for baseline weather data were pointed out by Belcher et al. (2005). First, the baseline data should represent the weather conditions for that period. Ideally, the baseline data is averaged data for a period of 30 years (Brisson et al. 2014). Second, as the climate change signal is expressed against a certain period in time, the baseline data used should be representative of the same period. If these periods do not correspond, the generated weather data can under- or over- estimate the climate change impact as pointed out by Kolokotroni et al. (2012). Third, as the climate change signal is mostly expressed as a mean change signal, it is important to use averaged weather data as baseline to avoid under- or overestimation. The morphing methodology is used among others by Chan (2011), Kolokotroni et al. (2012), CIBSE (2005) and Ferrari et al. (2008). In the UK, the Climate Change Weather Generator tool (CCWeatherGen tool or CCWorldWeatherGen for outside the UK) was created based on this morphing methodology making use of the HadCM31 GCM for the A2 emission scenario2 (Jentsch et al. 2008, 2013). Multiple studies moreover used this tool for climate change impact studies (Farrou et al. 2014; Kolokotroni et al. 2012; Roetzel and Tsangrassoulis 2012). An advantage of the morphing methodology is the low computational time. Various climate change scenarios can hence easily be applied. An important drawback of the morphing method is that the future weather data have the same character and variability as the current weather data (Belcher et al. 2005) and hence any changes in frequency of (extreme) weather events, although expected (Wouters et al. 2017; Kovats et al. 2014), are not considered. In addition, the mathematical basis of the method results in a limited representation and parametrization of climate physics possibly leading to meteorological inconsistencies (Belcher et al. 2005). This mostly leads to a limited representation of extreme weather events (Herrera 2017). Jentsch et al. (2013) furthermore point out that morphing data with GCMs tends to underestimate the climate change and they recommend to use RCM data for morphing if possible.

6.2.2 Weather Datasets

6.2.2.1 Multi-year Datasets

As mentioned before, climate models are typically multi-year datasets, usually covering periods of 30 years. Also, weather generators can result in multi-year datasets (Eames et al. 2011). The advantage of longer periods of data is the like- lihood that they are covering both typical and extreme weather conditions for that

1Hadley Centre Coupled Model, version 3. 2Medium-high emission scenario. One of the scenarios before the development of RCP scenarios to be used in GCMs for climate change. 6 Future Weather Data for Dynamic Building Energy Simulations … 119 period (Brisson et al. 2014). Such multi-year (30) datasets result in a high computational time if used in building simulations. In order to reduce the computational time, weather data files for selected years can be developed. In building energy simulation, very often 1-year weather data files are used, representing typical years or extreme years. Additionally, a combination of several 1-year weather data files can be used. This is, for example, proposed by Nik (2016) who suggests to use a combination of a typical year, an extreme cold and an extreme warm year. The various types of weather data files for a selected number of years are discussed in the subsequent sections.

6.2.2.2 Typical Years

Typical years are fictive years consisting of representative typical months (Barnaby and Crawley 2011) which are selected by comparing the distribution of each month with the long-term distribution of that month for the observations available (the Finkelstein–Schafer statistics 1971). When defining a typical year, various meteorological variables can be focused on and various relative importances (weighting factor) can be used for those variables. Various methods (Nik 2016; National Climate Data Center and U.S. Department of Commerce 1976; Kershaw et al. 2010; International Organization for Standardization 2005; Didier and Alfred 2002; Huang et al. 2014; Stoffel and Rymes 1998) exist within this type of weather data, and lead to various weather data files, such as Test Reference Year (TRY) (European Commission 1985), International Weather year for Energy Calculations (IWEC) (Huang et al. 2014; ASHRAE 2002; Thevenard and Brunger 2002) or a Typical Meteorological Year (TMY) (Didier and Alfred 2002). Typical years are often used to estimate the average energy use of a building (Barnaby and Crawley 2011). Typical years can either be based on observational data or data from a climate model. If these are based on observational data, the period for which the weather files are created depend on the available observational data for that specific location. Moreover, the locations for which such weather files can be derived are restricted to locations where observational data are available. Nevertheless, it is possible to create typical year data for locations without observational data by means of interpolation of the observed data of several relevant other locations. This is, for example, the case for the Meteonorm Typical Years (Meteotest 2017). If typical years are derived from climate model datasets, the amount of typical years for a certain region depend on the spatial resolution of the climate model. The Typical Downscaled Year (TDY) developed by Nik (2017) is an example of a typical year based on climate model data. For Belgium, the existing typical years (Test Reference Year or TRY (European Commission (EC) 1985), International Weather year for Energy Calculations or IWEC (ASHRAE 2002; Thevenard and Brunger 2002) and Meteonorm Typical Years (2017) are based on observational data. These cover at least one decade (Levermore and Doylend 2002; Lee et al. 2010) and are available for three locations 120 D. Ramon et al.

Table 6.1 Selection of typical years used in building simulation Acronym Complete name Weather variables + weights (in %) TRYa Test Reference Year (National Climate First step equal weighting of T, R and Data Center and U.S. Department of humidity; second step Wspd Commerce 1976; Kershaw et al. 2010; International Organization for Standardization 2005) a TMY Typical Meteorological Year (Didier Tmin (5%), Tmax (5%), Tmean (30%), and Alfred 2002) TDmin (2.5%), TDmax (2.5%), TDmean (5%), Wspdmax (5%), Wspdmean (5%), R (40%) a IWEC International Weather Year for Energy Tmin (5%), Tmax (5%), Tmean (30%), Calculations (Huang et al. 2014; TDmin (2.5%), TDmax (2.5%), TDmean ASHRAE 2002; Thevenard and (5%), Wspdmax (5%), Wspdmean (5%), R Brunger 2002) (40%) TDYb Typical Downscaled Year (Nik 2016)T WYECa Weather Year for Energy Calculations T, TD, R, precipitation (Stoffel and Rymes 1998; Crawley 1998) Notes (1) T = dry-bulb temperature, TD = dew-point temperature, Wspd = wind speed, R = radiation (2) Weather variables and related weighting factors changed over time for some of the methods and/or differ for different locations (3) aMethodology originally based on observations (4) bMethodology originally based on climate models

(Ostend, Uccle and Saint-Hubert) with the exception of the Meteonorm data. The Meteonorm software moreover allows to extract a typical year applying different methodologies (i.e. TMY2, TRY DWD, TRY DWD 1.1 (Deutscher Wetterdienst 2017) and TMY3 (Wilcox and Marion 2008)). A selection of typical year methods used in building simulations is summarized in Table 6.1. The table provides an overview of the meteorological variables considered in the basic dataset to define the typical year. Originally, the TRY (first one in the table) consisted of a selected single year out of a period of observations (National Climate Data Center and U.S. Department of Commerce 1976) and hence not of a combination of representative typical months from different years. Later, it evolved to a compilation of representative typical months, similar to the other methods in the table. For some of the methods, the considered variables and weighting scheme moreover evolved over the years (e.g. Typical Meteorological Year 2 or TMY2 weather format is adjusted from the TMY format adapting the weighting of the dry-bulb temperature and humidity (Marion and Urban 1995)). The variables considered moreover sometimes differ within one method in order to account for the most important variables in a specific location (e.g. UK-TRY and ISO-TRY) (National Climate Data Center and U.S. Department of Commerce 1976; Kershaw et al. 2010; Eames et al. 2015). Finally, an evolution of methods has been noticed due to the (non-)availability of solar data or more complex solar models to 6 Future Weather Data for Dynamic Building Energy Simulations … 121 generate solar data over time (Didier and Alfred 2002; Huang et al. 2014) (e.g. TMY2 included a more complex solar model than TMY (Marion and Urban 1995)).

6.2.2.3 Extreme Years

To assess the robustness of buildings to extreme weather conditions, to size HVAC systems or to perform climate uncertainty studies, insights into the variability of the climate and the occurrence of more extreme weather events are needed (Berger et al. 2014; Kershaw et al. 2011). An extreme or near-extreme year can be used to assess the building for these purposes. Currently, extreme years are 1-year weather data files which either include an extreme summer (e.g. Design Summer Year (DSY)) or winter (e.g. Extreme Cold Year (ECY)) or combination of both (e.g. Extreme Meteorological Year (XMY)). There are several ways to select/compose an extreme year. The first approach uses a similar methodology as for defining typical years (i.e. making use of the Finkelstein–Schafer statistics). Instead of searching for the most representative typical months, the most deviating months are selected or a certain percentile in the distribution is selected (Nik 2016; Weather et al. 2002). For example, an XMY is defined by using the same approach as for defining a TMY. Although the approaches to define an XMY and TMY consider the same meteorological variables and weighting scheme (Ferrari and Lee 2008), this is not always the case. The considered meteorological variables and related weighting factors for defining extreme years sometimes differ from the ones used to define a typical year (e.g. an Untypical Meteorological Year applies the same methodology but different weighting than the Weather Year for Energy Calculations version 2 (Narowski et al. 2013; Stoffel and Rymes 1998)). The Meteonorm software allows to select a P10 (minima) or P90 (maxima) year which has the probability to happen once a decade (Meteotest 2017). The second approach selects the extreme year based on a calculated value. The DSY calculates, for example, the average dry-bulb temperature for the sum- mertime (Weather et al. 2002) while a Hot Summer Year (HSY) calculates the Weighted Cooling Degree-Hours3 (WCDH) (Liu et al. 2016; CIBSE 2014) to select a year. The use of one calculated value has some limitations. The DSY, for instance, represents a year with a warm summer based on the average monthly temperature for April to September and risks to neglect a summer with critical heatwave periods if the average temperatures are lower for the latter (CIBSE 2009, 2014). It should furthermore be noted that an extreme year does not provide insights into the average discomfort to be expected on yearly basis. For example, a DSY is an extreme warm year to be expected to happen once each 8 years (CIBSE 2009).

3Deﬁned as “the cumulative squared hourly difference between the outdoor dry-bulb temperature and the adaptive thermal comfort temperature” (CIBSE 2014: p. 1). 122 D. Ramon et al.

Table 6.2 Selection of extreme weather years used in building simulation Acronym Complete name Weather variables + weights (in %) DSYa Design Summer Year Average T for period April–September (CIBSE 2014) a XMY Extreme Meteorological Tmin (5%), Tmax (5%), Tmean (30%), TDmin (2.5%), Year (Ferrari and Lee 2008) TDmax (2.5%), TDmean (5%), Wspdmax (5%), Wspdmean (5%), R (40%) (based on TMY) UMYa Untypical Meteorological T, R, Wspd (based on WYEC) Year (Narowski et al. 2013) HSYa Hot Summer Year (Liu et al. Highest WCDH (HSY-1) or most hours of PETc 2016) over 23 °C (HSY-2) EWYb Extreme Warm Year (Nik T 2016) ECYb Extreme Cold Year (Nik T 2016) DRYa Design Reference Year T, R, humidity, wind speed (weather generator (Watkins et al. 2013) combined with weighting scheme) P10/P90a P10 (cold)/P90 (warm) T, R extreme year (Meteotest 2017) Notes (1) T = dry-bulb temperature, TD = dew-point temperature, Wspd = wind speed, R = radiation (2) WDCH = weighted cooling degree-hours, PET = physiologically equivalent temeprature (Höppe 1999) (3) Weather variables and related weighting factors changed over time for some of the methods and/or differ for different locations (4) aMethodology originally based on observations (5) bMethodology originally based on climate models (6) cPhysiologically equivalent temperature (Höppe 1999)

Similar as with typical years, extreme years can be created based on observational data, climate models or weather generator output. For Belgium, the existing extreme years are based on observations (the Meteonorm P10 and P90 years for the locations of Uccle, Ostend and Saint-Hubert). A selection of extreme weather years which are commonly used in building simulations is summarized in Table 6.2.

6.2.2.4 Representative Datasets

The above described typical and extreme years are combined by Nik (2016)in so-called representative datasets. Nik synthesizes three years out of a multi-year period, extracting a typical and an extreme cold and warm year from RCMs (multiple if available). A typical downscaled year (TDY) is extracted representing the typical conditions of the full period. Further, the extreme cold and extreme warm years are extracted representing, respectively, coldest and warmest conditions from the considered period. To derive these three years, a similar methodology is used as the one for the TMY, making use of the Finkelstein–Schafer statistics. In this case, 6 Future Weather Data for Dynamic Building Energy Simulations … 123 temperature is considered as variable to select the most extreme/typical months. This approach moreover allows to consider one or several future climate scenarios.

6.2.3 Recommendations for the Use of Future Weather Data for Building Energy Simulations

When aiming at designing climate resilient buildings, various aspects can be investigated, ranging from the average energy consumption and thermal comfort to resilience in extreme weather conditions. Depending on the goal of the study, the most appropriate weather data need to be selected. When investigating the average energy consumption and thermal comfort, typical years are preferred, while to ensure resilience in extreme weather conditions, multi-year datasets or extreme weather years are better suited. It is furthermore recommended that multi-year datasets, whether used directly or used for deriving extreme weather years from them, should cover a sufficiently long period (typically 30 years) to include the variability of the climate. When extreme weather data files are used it is furthermore important to select the appropriate one (s) for the goal of the study as some files rather represent years with high summer temperatures and not necessarily years with higher occurrence of extreme weather events, or the other way around. As discussed in Sect. 6.2.1, various methods exist to create future weather data from which then typical or extreme years can be derived. Each of the existing methods has its strong points and limitations. It is hence not only important to select the most appropriate weather data file (typical versus extreme, or a combination of both), but also to select the most appropriate method used to create the future weather data, in line with the goal of the study. While morphing is suited to investigate the average energy performance in a future climate realization, the fact that the baseline data is averaged data and climate change signal a mean change makes it less suited for the assessment in extreme conditions. Dynamically downscaled climate models can be used for both average as extreme assessment purposes. In particular, when downscaled to convection-permitting scale (<4 km), these models are interesting to be used for estimating the average energy consumption and thermal comfort in average and extreme situations. These models have an improved performance of the physical process, urban effects (e.g. urban heat island effect) and climate change signal. However, they only represent one future climate realization. As climate change is inherently characterised by uncertainties, considering different possible climate realizations is of major importance to design robust buildings. When the research aims at addressing uncertainties, probably the weather generators are the most appropriate. Weather generators are able to produce various climate realizations based on several climate change scenarios. Nevertheless, their spatial resolution is often coarser and more important 124 D. Ramon et al. meteorological variables are based on their statistical and mathematical relations which could result in meteorological inconsistencies. Research focusing on building energy simulations with future climate data in the Belgian context is rare. In 2015, a convection-permitting climate model for Belgium for the end of the century became available. A 30-year weather data file was derived from this model to be used for building energy simulations (Ramon et al. 2017). However, the 30-year data resulted in high computational time. Therefore, a synthesized dataset has been derived from this 30-year dataset representing a typical and extreme cold and warm year. In Sect. 6.3, this synthesized dataset (TDY, ECY and EWY) is presented and compared with other typical and extreme years (TMY and DSY) extracted from the 30-year dataset. Comparisons are made for both the current and future climate perspective. The new and existing datasets are moreover used in a simple case study to investigate differences in results obtained.

6.3 Future Weather Data for the Belgian Context, Applied to an Ofﬁce Building

6.3.1 Building Energy Model

A cellular office from a representative Flemish office building is simulated for this study making use of the dynamic energy simulation program EnergyPlus v8.7. The building has a concrete structure and a high insulation level (average U-value of 0.42 W/m2 K) (Belcher et al. 2005). A more detailed overview of the building elements is given in Table 6.3. A cellular office room located in the middle of the building is chosen for the simulations (Fig. 6.1). The temperature of internal walls, floors and ceilings is assumed to be the same at both sides of those elements. Heat can be accumulated in those elements and dissipated later, but there is no resulting flow. This is only a good assumption if the use of the adjacent rooms is similar to

Table 6.3 Building element composition Building Element composition Total U-value element [W/m2/K] External wall Facing brick–air layer–PIR insulation 0.14 15 cm–concrete panel 20 cm Internal wall gypsum plasterboard 2.5 cm–metal stud with n/a mineral wool 5 cm–gypsum plasterboard 2.5 cm Internal ﬂoor suspended ceiling–air layer–prestressed concrete n/a 35 cm–cement screed 9 cm–PVC Tiles 1 cm Windows aluminium frame–double-glazed (g-value 0.275) 1.4 6 Future Weather Data for Dynamic Building Energy Simulations … 125

Fig. 6.1 Simulation setup: schematic building and considered office room the room studied, leading to only small temperature differences between the rooms. To study the effect of orientation, the model is run for four orientations, the glazed façade facing north, east, south and west. The schedules for occupancy, lighting and equipment in the office room are modelled based on the NCM modelling guide (Communities & Local Government 2008). The number of people per zone is defined in terms of area per person based on the architectural plans with a metabolic rate of 117 W/person (Architectenvenootschap ar-te bcvba: BelOrta; ASHRAE 2009). Lights are modelled based on the lighting level (Watts) installed in the building. The office equipment is set to a load of 11.77 W/m2 and a radiant fraction of 0.2 (ASHRAE 2009). For the HVAC system, an ideal air loads system with unlimited capacity is implemented to get insights into the increase/decrease of energy to keep the room comfortable. Because of the unlimited capacity, heating and cooling are performed till the set points are met. The operative temperature (average of air temperature and surface temperature of walls, ceiling and floor) is used to control the ideal air loads system. Heating and cooling are assumed all year long. The heating set point is set on 22 °C during occupancy and 15 °C for other moments. Cooling is set to 24.5 °C during the day and 35 °C during the night and weekend for the summer period and heating season.

6.3.2 Description and Analysis of Weather Data Used

Several weather ﬁles were extracted for Uccle, centrally located in Belgium (120 km from coast, 104 m above sea level). First, two EC-Earth4 driven Convection-Permitting Models (CPMs) are used (Hazeleger et al. 2010). The CPM uses COSMO-CLM5 (version 5.0) as regional

4EC-Earth is a global climate model developed by a European consortium (Hazeleger et al. 2010). 5Consortium for small scale modelling in Climate mode is a regional climate model (Vanden Broucke 2017). 126 D. Ramon et al. climate model. A two-step nesting strategy is used to obtain the model with a 2.8 km horizontal resolution for the Belgian domain. In the first step, the 12 km resolution RCM simulation was nested in the EC-Earth domain, and in the second step, the 2.8 km resolution simulation was nested in the 12 km simulation (Vanden Broucke 2017). To incorporate the local characteristics and the three-dimensional structure of the urban canopy, the CPM model is extended with the urban land-surface scheme TERRA_URB (v2.0) making use of the SURY (Semi-empirical URban canopY) parametrization (Wouters et al. 2016). TERRA_URB calculates radiation, heat and moisture fluxes between the urban environment and the atmosphere based on the bulk parameters (e.g. heat conductivity, albedo, aerodynamical roughness length, etc.). In addition, SURY translates the urban canopy parameters (e.g. building height, roof fraction, albedo, heat conductivity, etc.) into bulk parameters used in the climate models (Wouters et al. 2016). The model systematically evaluated and found to reproduce both the observed coarse temperature and the urban heat islands of the study domain very well in general, and the hourly and daily variability in particular (Wouters et al. 2017). From the model simulation, an epw file is extracted for the 30-year period 1975–2004 (further referred to as EC-Earth) and the 30-year period 2069–2098 (further referred to as EC-Earth Future) based on the methodology described in (Ramon et al. 2017). The first year is each time used as a spin-up for the simulations. For the future time period, the climate change scenario chosen is one of the 16 EC-Earth members specified for the RCP8.5 scenario which has the most median climate change signal regarding temperature, precipitation and widespread circulation patterns for Belgium (IPCC 2013; Vanden Broucke 2017). Figure 6.2 shows the temperature distribution for both periods. It is clearly visible that the mean monthly temperature increases over the full year (with an average increase of 2.9 °C over the 12 months). For July, the increase in the mean temperature even reaches a difference of 3.8 °C. The increase in the monthly minimum and maximum temperature is slightly more extreme than the increase of the monthly mean temperature. For August, the monthly maximum temperature increases even with 4.9 °C towards the future. For February, the monthly minimum temperature increases with 5.2 °C. Figure 6.3 shows an increase in radiation during summer (May- September) and a decrease for the rest of the year (October–March) with exception of November. The increase and decrease are caused by a change in the direct component of the radiation, while the diffuse component remains the same. Second, some 1-year weather files are generated from the climate models for EC-Earth and EC-Earth Future. A Typical Meteorological Year (TMY) and a Typical Downscaled Year (TDY) are selected as typical years. The TMY methodology selects per month the most typical year from the climate model data based on the Finkelstein–Schafer statistics. Radiation and dry-bulb temperature each get a contribution of 40% in the weighting of the climate variables, wind speed and dew-point temperature a contribution of 10% each (Huang et al. 2014). The TDY methodology applies the same Finkelstein–Schafer statistics, but only based on dry-bulb temperature (Nik 2016). 6 Future Weather Data for Dynamic Building Energy Simulations … 127

Fig. 6.2 Distribution of monthly dry-bulb temperature EC-Earth (1976–2004—red) and EC-Earth Future (2070–2098—black)

Fig. 6.3 Distribution of monthly global radiation for EC-Earth (1976–2004—red) and EC-Earth Future (2070–2098—black)

As the TDY, ECY and EWY only use dry-bulb temperature as variable to determine the representativeness of the months, the good ﬁt of these years with the mean and extreme temperatures of the 30-year period is expected (Fig. 6.4). Although the TMY considers both radiation and temperature to select the most representative months, some deviations are visible for the temperature. For radiation, the TMY has a good ﬁt with the median values over the 30-year period. In 128 D. Ramon et al.

Fig. 6.4 Temperature distribution for EC-Earth (1976–2004) with monthly mean Typical Meteorological Year (TMY), Typical Downscaled Year (TDY), Extreme Warm Year (EWY), Extreme Cold Year (ECY) and Design Summer Year (DSY) general, the TDY also has a good ﬁt with the 30-year period (Fig. 6.5), however with some values falling out of the range of the ﬁrst to third quartile of the 30-year distribution (e.g. October). As the TDY is only selected on temperature, some deviations were expected. The ECY and EWY both have monthly values for radiation within the upper and lower percentiles of the 30-year distribution for the EC-Earth run. However, for EC-Earth future run (Fig. 6.6), both the ECY and EWY have monthly global radiation values close to the extremes of the 30-year distribution (e.g. September for ECY, October for EWY).

6.3.3 Energy Demand

In this simulation, an ideal air loads system with an unlimited capacity is used. Consequently, the temperature set points are always met and the energy demand for heating and cooling gives an idea of the energy need for the thermal comfort for that period (e.g. higher cooling demand for months with higher temperatures). For the monthly heating load, Fig. 6.7 shows that the heating load calculated with the TDY corresponds well with the median monthly heating load for the 30-years of the climate model. For the TMY, the heating load often falls out of the 25–75th percentile of the 30-year calculation. Further, the ECY gives a good representation of the maximum heating load of the 30-year calculation and the EWY for the minimum. These observations also hold true for the other orientations and EC-Earth Future simulation. 6 Future Weather Data for Dynamic Building Energy Simulations … 129

Fig. 6.5 Global radiation distribution for EC-Earth (1976–2004) with monthly mean Typical Meteorological Year (TMY), Typical Downscaled Year (TDY), Extreme Warm Year (EWY), Extreme Cold Year (ECY) and Design Summer Year (DSY)

Fig. 6.6 Global radiation distribution for EC-Earth Future (2070-2098) with monthly mean Typical Meteorological Year (TMY), Typical Downscaled Year (TDY), Extreme Warm Year (EWY) and Extreme Cold Year (ECY) 130 D. Ramon et al.

Fig. 6.7 Distribution of total monthly heating load EC-Earth (1976–2004) and total monthly heating load for Typical Meteorological Year (TMY), Typical Downscaled Year (TDY), Extreme Warm Year (EWY), Extreme Cold Year (ECY) and Design Summer Year (DSY)

For the monthly cooling load, the TDY and TMY result in loads mostly within the 25–75th percentile of the 30-year climate model results (Figs. 6.8 and 6.9). For both the EC-Earth and EC-Earth Future, the DSY tends to underestimate the maximal cooling load. This is caused by the methodology to select a DSY. A full year is selected based on the average temperature from April to September. Consequently, an on average warm year is selected, however not necessarily a year with a lot of heatwave periods, which is the case for Uccle. These can also occur in a summer with a lower average temperature. Therefore, these years can be missed in the selection of the DSY. Looking to the ability of the EWY to represent the maximum cooling load, an overall good ﬁt is observed. In some cases (e.g. September in the EC-Earth simulation), the EWY underestimates the maximum cooling load even though a month with a higher average temperature is selected for the EWY. Moreover, in the September case of the EC-Earth simulation this month even had some days ful- ﬁlling heatwave requirements (minimum temperature above 18.2 °C and maximum temperature above 29.6 °C (Wouters et al. 2017)). In this case, the lower cooling need is caused by some days with a very low temperature and in addition lower irradiation on the façade. The ECY represents the minimum cooling load overall well. However, for the EC-Earth Future (Fig. 6.9), a higher cooling load than expected (e.g. September) is often reported. For September, this can be explained by the high radiation in the ECY for that month (Fig. 6.6) mainly caused by the direct component of the radiation. Hence, the solar heat gains through the windows increase and cause 6 Future Weather Data for Dynamic Building Energy Simulations … 131

Fig. 6.8 Distribution of total monthly cooling load for EC-Earth (1976–2004) and total monthly cooling load for Typical Meteorological Year (TMY), Typical Downscaled Year (TDY), Extreme Warm Year (EWY), Extreme Cold Year (ECY) and Design Summer Year (DSY)

Fig. 6.9 Distribution of total monthly cooling load for EC-Earth Future (2070–298) and total monthly cooling load for Typical Meteorological Year (TMY), Typical Downscaled Year (TDY), Extreme Warm Year (EWY), Extreme Cold Year (ECY) and Design Summer Year (DSY) 132 D. Ramon et al. the higher cooling need. A similar explanation is found for October; however, the monthly radiation is lower which results in a lower resulting cooling load. In general, the 30-year climate models have the advantage to represent the climate variability over a longer period. A combination of 1-year weather files (TDY, EWY and ECY) as proposed by Nik (2016) to reduce computational time proved to capture the average climate over the 30-years as well as the extreme climate events for the current and future climate conditions. However, some deviations for the ECY in the EC-Earth Future run were found. For the average heating and cooling load, differences are identified between the different 1-year weather files (TDY and TMY) where the TDY results in a better agreement with the median 30-year loads for current and future climate conditions. The consideration of radiation in the TMY seems less important for Uccle. For the extreme cooling load, the DSY leads to an underestimation of the maximal cooling load over 30-years caused by the selection based on an average temperature for April till September. Extreme heatwave periods can happen during an on average cool summer.

6.4 Conclusion, Critical Reﬂections and Outlook

This chapter provides an overview of common methods to create future weather data and widely used weather datasets which can be used to assess the energy and thermal comfort performance of buildings under climate change for Belgium. In addition, recently developed future weather data files for Belgium were used to analyse the energy performance of a representative Flemish office room. The future weather data were generated from a recently developed convection-permitting climate model (CPM) for Belgium and consist of a Typical Meteorological Year (TMY), Typical Downscaled Year (TDY), Design Summer Year (DSY), Extreme Warm Year (EWY) and Extreme Cold Year (ECY). For reasons of representation, a CPM for both the current and future climate conditions was used in the case study. The results obtained were moreover compared with results retrieved when using the existing future weather datasets. Based on the analysis of the advantages and disadvantages of the methods to create future weather data, climate models and weather generators are identified as the most preferred for the estimation of the average energy consumption and thermal comfort in average and extreme situations. Climate models have the advantage to provide insights into future climate conditions with a meteorological consistency. In particular, Regional Climate models downscaled to a convection-permitting scale (CPMs with a resolution <4 km) allow to represent climate differences due to territorial settings (e.g. city versus suburban versus rural area) and better represent extreme weather events. The drawback of these models is however that they represent only one climate realization. Weather generators are interesting as they can generate multiple climate realizations with a limited computational time. The use of statistical relations in this methodology, however, can 6 Future Weather Data for Dynamic Building Energy Simulations … 133 lead to meteorological inconsistencies. Further research is needed to investigate how both methodologies can be combined to keep the detailed representation of local climate processes and combine this with climate change uncertainties. Looking at different weather datasets, 1-year weather files represent average or extreme weather conditions (respectively, typical and extreme weather years). To capture the climate variability, these files need to be combined in so-called synthesized weather data files. This was illustrated for the Belgian situation by combining a TDY with an EWY and ECY for the location of Uccle, as proposed in the methodology of Nik (2016). This resulted in an overall good representation of the energy need for heating and cooling in average and extreme weather conditions. The influence of the methodological choices to extract these 1-year weather files from the 30-year climate data is highlighted as different results were obtained when different meteorological variables were considered for the creation of the 1-year files. Differences in average energy need were, for example, obtained when using TDY (created based on dry-bulb temperature) and a TMY (selected based on dry-bulb temperature, radiation, wind speed and humidity). Furthermore, the use of a DSY (full year selected based on average temperature April–September) underestimates extreme warm weather events and leads to an underestimation of the energy need for cooling compared to the median cooling load for the 30-years of the climate model. Finally, an ECY, which is expected to provide insights into the minimum cooling loads, led to higher cooling loads than expected based on the 30-years climate model. As the selection of an extreme cold month in the ECY weather data file is only based on temperature, still high radiation values can occur leading to a higher cooling need.

Acknowledgements This chapter is part of an SBO Ph.D. fellowship ‘Towards future-proof buildings in Flanders: Climate and Life Cycle modelling for resilient ofﬁce buildings’ (1S97418 N) funded by Research Foundation Flanders (FWO).

References

Adelard L, Boyer H, Garde F, Gatina J, Adelard L, Boyer H, Garde F, Detailed JG (2012) Detailed weather data generator for building simulations to cite this version: HAL Id: hal-00765831 a detailled weather data generator for buildings simulation, vol 31, 75–88 Andrić I, Pina A, Ferrão P, Fournier J, Lacarrière B, Le Corre O (2017) The impact of climate change on building heat demand in different climate types. Energy Build 149. https://doi.org/ 10.1016/j.enbuild.2017.05.047 Architectenvenootschap ar-te bcvba (2015) Architectural plans and model BelOrta ofﬁce building ASHRAE (2002) Brussels 064510 (IWEC) ASHRAE (2009) ASHRAE handbook-fundamentals, pp 21.1–21.67. https://doi.org/10.1017/ cbo9781107415324.004 ASHRAE (2017) Chapter 14, Climatic design information. In: ASHRAE handbook-fundamentals Barnaby CS, Crawley DB (2011) Weather data for building performance simulation. In: Building performance simulation for design and operation Belcher S, Hacker J, Powell D (2005) Constructing design weather data for future climates. Build Serv Eng Res Technol 26:49–61. https://doi.org/10.1191/0143624405bt112oa 134 D. Ramon et al.

Berger T, Amann C, Formayer H, Korjenic A, Pospischal B, Neururer C, Smutny R (2014a) Impacts of climate change upon cooling and heating energy demand of office buildings in Vienna, Austria. Energy Build 80:517–530. https://doi.org/10.1016/j.enbuild.2014.03.084 Berger T, Amann C, Formayer H, Korjenic A, Pospichal B, Neururer C, Smutny R (2014b) Impacts of urban location and climate change upon energy demand of office buildings in Vienna. Austria Build Env 81:258–269. https://doi.org/10.1016/j.buildenv.2014.07.007 Brisson E, Demuzere M, Willems P, van Lipzig NPM (2014) Assessment of natural climate variability using a weather generator. Clim Dyn 44:495–508. https://doi.org/10.1007/s00382- 014-2122-8 Brisson E, Van Weverberg K, Demuzere M, Devis A, Saeed S, Stengel M, van Lipzig NPM (2016) How well can a convection-permitting climate model reproduce decadal statistics of precipitation, temperature and cloud characteristics? Clim Dyn 47:3043–3061. https://doi.org/ 10.1007/s00382-016-3012-z Brotas L, Nicol F (2017) Architectural science review estimating overheating in European dwellings. Archit Sci Rev 603:180–191. https://doi.org/10.1080/00038628.2017.1300762 Chan ALS (2011) Developing future hourly weather files for studying the impact of climate change on building energy performance in Hong Kong. Energy Build 43:2860–2868. https:// doi.org/10.1016/j.enbuild.2011.07.003 Chow DHC, Levermore GJ (2010) The effects of future climate change on heating and cooling demands in office buildings in the UK. Build Serv Eng Res Technol 31:307–323 CIBSE (2002) CIBSE guide J: weather, solar and illuminance data. Chart Inst Build Serv Eng 455 CIBSE (2005) TM36: climate change and the indoor environment: impacts and adaptation. Charted Inst Build Serv Eng CIBSE (2009) Use of climate change scenarios for building simulation: the CIBSE future weather years Use of climate change scenarios for building simulation: the CIBSE future weather years CIBSE (2014) Design summer years for London TM49 Communities & Local Government (2008) National Calculation Methodology (NCM) modelling guide (for buildings other than dwellings in England and Wales). Communities local government, pp 1–34 Crawley D (1998) Which weather data should you use for energy simulations of commercial buildings? Am Soc Heat Refrig Air-Cond 1–18 Crawley DB (2008) Estimating the impacts of climate change and urbanization on building performance. J Build Perform Simul 1:91–115. https://doi.org/10.1080/19401490802182079 Deutscher Wetterdienst (2017) Handbuch Ortsgenaue Testreferenzjahre von zukünftige Witterungsverhältnisse 45 Didier J, Alfred P (2002) The development of typical weather years for international locations: part I, algorithms Eames M, Kershaw T, Coley D (2011) On the creation of future probabilistic design weather years from UKCP09. Build Serv Eng Res Technol 32:127–142. https://doi.org/10.1177/ 0143624410379934 Eames ME, Ramallo-Gonzalez AP, Wood MJ (2015) An update of the UK’s test reference year: the implications of a revised climate on building design. Build Serv Eng Res Technol 37:316– 333. https://doi.org/10.1177/0143624415605626 EnergyPlus (2016) EnergyPlusTM Version 8.7 documentation: auxiliary programs European Commission (EC) (1985) Test Reference Years TRY, Weather data sets for computer simulations of solar energy systems and energy consumption in buildings. Brussels, Belgium Farrou I, Kolokotroni M, Santamouris M (2014) Building envelope design for climate change mitigation: a case study of hotels in Greece. Int J Sustain Energy 1–24. https://doi.org/10.1080/ 14786451.2014.966711 Ferrari D, Lee T (2008) Beyond Tmy: climate data for specific applications. Sol Energy 1–12 Finkelstein JM, Schafer RE (1971) Improved goodness-of-fit tests. Biometrika 58:641–645. https://doi.org/10.1093/biomet/58.3.641 6 Future Weather Data for Dynamic Building Energy Simulations … 135

Hazeleger W, Severijns C, Semmler T, Ştefǎnescu S, Yang S, Wang X, Wyser K, Dutra E, Baldasano JM, Bintanja R, Bougeault P, Caballero R, Ekman AML, Christensen JH, Van Den Hurk B, Jimenez P, Jones C, Kållberg P, Koenigk T, McGrath R, Miranda P, Van Noije T, Palmer T, Parodi JA, Schmith T, Selten F, Storelvmo T, Sterl A, Tapamo H, Vancoppenolle M, Viterbo P, Willén U (2010) EC-Earth: a seamless Earth-system prediction approach in action. Bull Am Meteorol Soc 91:1357–1363. https://doi.org/10.1175/2010BAMS2877.1 Herrera M, Natarajan S, Coley DA, Kershaw T, Ramallo-González AP, Eames M, Fosas D, Wood M (2017) A review of current and future weather data for building simulation. Build Serv Eng Res Technol. https://doi.org/10.1177/0143624417705937 Holmes MJ, Hacker JN (2007) Climate change, thermal comfort and energy: meeting the design challenges of the 21st century. Energy Build 39:802–814. https://doi.org/10.1016/j.enbuild. 2007.02.009 Höppe P (1999) The physiological equivalent temperature—a universal index for the biomete- orological assessment of the thermal environment. Int J Biometeorol 43:71–75. https://doi.org/ 10.1007/s004840050118 Huang YJ, Su F, Seo D, Krarti M (2014) Development of 3012 IWEC2 weather files for international locations (RP-1477). ASHRAE Trans 120:340–355 Hutchinson MF (1987) Methods of generation of weather sequences. In: Bunting AH (ed) Agricultural environments: characterization, classification and mapping. CAB International, Wallingford, UK International Organization for Standardization (2005) 15927-4 hygrothermal performance of buildings—calculation and presentation of climatic data—part 4: hourly data for assessing the annual energy use for heating and cooling, Geneva IPCC (2013) Climate change 2013: the physical science basis. In: Contribution of working group I to the fifth assessment report of the intergovernmental panel on climate change. CUP, Cambridge, United Kingdom and New York, USA Jacob D, Petersen J, Eggert B, Alias A, Christensen OB, Bouwer LM, Braun A, Colette A, Déqué M, Georgievski G, Georgopoulou E, Gobiet A, Menut L, Nikulin G, Haensler A, Hempelmann N, Jones C, Keuler K, Kovats S, Kröner N, Kotlarski S, Kriegsmann A, Martin E, van Meijgaard E, Moseley C, Pfeifer S, Preuschmann S, Radermacher C, Radtke K, Rechid D, Rounsevell M, Samuelsson P, Somot S, Soussana JF, Teichmann C, Valentini R, Vautard R, Weber B, Yiou P (2014) EURO-CORDEX: new high-resolution climate change projections for European impact research. Reg Environ Chang 14:563–578. https://doi.org/10. 1007/s10113-013-0499-2 Jentsch MF, Bahaj AS, James PAB (2008) Climate change future proofing of buildings-generation and assessment of building simulation weather files. Energy Build 40:2148–2168. https://doi. org/10.1016/j.enbuild.2008.06.005 Jentsch MF, James PAB, Bourikas L, Bahaj AS (2013) Transforming existing weather data for worldwide locations to enable energy and building performance simulation under future climates. Renew Energy 55:514–524. https://doi.org/10.1016/j.renene.2012.12.049 Jones P, Harpham C, Kilsby C, Glenis V, Burton A (2010) UK climate projections science report: projections of future daily climate for the UK from the weather generator Kendon EJ, Ban N, Roberts NM, Fowler HJ, Roberts MJ, Chan SC, Evans JP, Fosser G, Wilkinson JM (2017) Do convection-permitting regional climate models improve projections of future precipitation change? Bull Am Meteorol Soc 98:79–93. https://doi.org/10.1175/ BAMS-D-15-0004.1 Kershaw T, Eames M, Coley D (2010) Comparison of multi-year and reference year building simulations. Build Serv Eng Res Technol 31:357–369. https://doi.org/10.1177/ 0143624410374689 Kershaw T, Eames M, Coley D (2011) Assessing the risk of climate change for buildings: a comparison between multi-year and probabilistic reference year simulations. Build Env 46:1303–1308. https://doi.org/10.1016/j.buildenv.2010.12.018 136 D. Ramon et al.

Kikumoto H, Ooka R, Arima Y, Yamanaka T (2015) Study on the future weather data considering the global and local climate change for building energy simulation. Sustain Cities Soc 14:404– 413. https://doi.org/10.1016/j.scs.2014.08.007 Kolokotroni M, Ren X, Davies M, Mavrogianni A (2012) London’s urban heat island: impact on current and future energy consumption in office buildings. Energy Build 47:302–311. https:// doi.org/10.1016/j.enbuild.2011.12.019 Kovats RS, Valentini R, Bouwer LM, Georgopoulou E, Jacob D, Martin E, Rounsevell M, Soussana JF (2014) Europe. In: Barros VR, Field CB, Dokken D, Mastrandrea M, Mach KJ, Bilir TE, Chatterjee M, Ebi KL, Estrada YO, Genova RC, Girma B, Kissel ES, Levy AN, MacCracken S, Mastrandrea PR, White L (eds) Climate change 2014: impacts, adaptation and vulnerability—contributions of the working group II to the fifth assessment report, pp 1267– 1326. Cambridge University Press, Cambridge, United Kingdom and New York, USA Lee K, Yoo H, Levermore GJ (2010) Generation of typical weather data using the ISO Test Reference Year (TRY) method for major cities of South Korea. Build Env 45:956–963. https:// doi.org/10.1016/j.buildenv.2009.10.002 Levermore GJ, Doylend NO (2002) North American and European hourly based weather data and methods for HVAC building energy analyses and design by simulation. ASHRAE Trans 108:1053–1062 Levermore G, Courtney R, Watkins R, Cheung HKW, Parkinson JB, Laycock P, Natarajan S, Nikolopoulou M, McGilligan C, Muneer T, Tham Y, Underwood CP, Edge JS, Du H, Sharples S, Kang J, Barclay M, Sanderson M (2014) Deriving and using future weather data for building design from UK climate change projections—an overview of the COPSE project, pp 1–7 Liu C, Kershaw T, Eames ME, Coley DA (2016) Future probabilistic hot summer years for overheating risk assessments. Build Env 105:56–68. https://doi.org/10.1016/j.buildenv.2016. 05.028 Marion W, Urban K (1995) User’s manual for TMY2s: derived from the 1961–1990 national solar radiation data base MBE KTN (2013) Guidance for making the case for climate change adaptation in the built environment Meteotest: Remund J, Müller S, Kunz S, Huguenin-Landl B, Studer C, Cattin R (2017) Meteonorm handbook part I: software, global meteorological database version 7 software and data for engineers, planers and education. http://www.meteonorm.com Moss R, Babiker M, Brinkman S, Calvo E, Carter T, Edmonds J, Elgizouli I, Emori S, Erda L, Hibbard K, Jones R, Kainuma M, Kelleher J, Lamarque JF, Manning M, Matthews B, Meehl J, Meyer L, Mitchell J, Nakicenovic N, O’Neill B, Pichs R, Riahi K, Rose S, Runci P, Stouffer R, van Vuuren D, Weyant J, Wilbanks T, van Ypersele JP, Zurek M (2008) Towards new scenarios for analysis of emissions, climate change, impacts and response strategies Narowski P, Janicki M, Heim D (2013) Comparison of untypical meteorological years (umy) and their influence on building energy performance simulations. In: Proceedings of BS 2013: 13th conference of the international building performance simulation association, pp 1414–1421 National Climate Data Center, U.S. Department of Commerce (1976) Test Reference Year. Tape Ref Man TD-9706 Nik VM (2016) Making energy simulation easier for future climate—synthesizing typical and extreme weather data sets out of regional climate models (RCMs). Appl Energy 177:204–226. https://doi.org/10.1016/j.apenergy.2016.05.107 Nik VM (2017) Application of typical and extreme weather data sets in the hygrothermal simulation of building components for future climate—a case study for a wooden frame wall. Energy Build 154:30–45. https://doi.org/10.1016/j.enbuild.2017.08.042 Nik VM, Sasic Kalagasidis A (2013) Impact study of the climate change on the energy performance of the building stock in Stockholm considering four climate uncertainties. Build Env 60:291–304. https://doi.org/10.1016/j.buildenv.2012.11.005 6 Future Weather Data for Dynamic Building Energy Simulations … 137

Nik VM, Coccolo S, Kämpf J, Scartezzini JL (2017) Investigating the importance of future climate typology on estimating the energy performance of buildings in the EPFL campus. Energy Procedia 122:1088–1093. https://doi.org/10.1016/j.egypro.2017.07.434 Prein AF, Langhans W, Fosser G, Ferrone A, Ban N, Goergen K, Keller M, Tölle M, Gutjahr O, Feser F, Brisson E, Kollet S, Schmidli J, Van Lipzig NPM, Leung R (2015) A review on regional convection-permitting climate modeling: demonstrations, prospects, and challenges. Rev Geophys 53:323–361. https://doi.org/10.1002/2014RG000475 Ramon D, Wouters H, Van Lipzig N, Allacker K (2017) Comparison of high-resolution climate model data with a Test Reference Year for building simulations. In: Brotas L, Roaf S, Nicol JF (eds) Proceedings of 33rd PLEA international conference, pp 1865–1872. NCEUB 2017 RICS (2015) Climatic risk toolkit. The impact of climate change in the non-domestic real estate sector of eight European countries Roetzel A, Tsangrassoulis A (2012) Impact of climate change on comfort and energy performance in offices. Build Env 57:349–361. https://doi.org/10.1016/j.buildenv.2012.06.002 Shen P (2017) Impacts of climate change on U.S. building energy use by using downscaled hourly future weather data. Energy Build 134:61–70. https://doi.org/10.1016/j.enbuild.2016.09.028 Shibuya T, Croxford B (2016) The effect of climate change on office building energy consumption in Japan. Energy Build 117. https://doi.org/10.1016/j.enbuild.2016.02.023 Stoffel TL, Rymes MD (1998) Production of the weather year for energy calculations version 2 (WYEC2) data files. ASHRAE Trans 487–497 Struck C, de Wilde P, Evers J, Hensen JLM, Plokker W (2009) On selecting weather data sets to estimate a building design’s robustness to climate variations. In: Proceedings of the 11th IBPSA building simulation conference, pp 513–520 Thevenard DJ, Brunger AP (2002) The development of Typical Weather Years for international locations: part II, production. ASHRAE Trans 480–486 Van Paassen AHC, Luo QX (2002) Weather data generator to study climate change on buildings. Build Serv Eng Res Technol 23:251–258. https://doi.org/10.1191/0143624402bt048oa Vanden Broucke S, Wouters H, Demuzere M, van Lipzig NPM (2017) Added value of convection-permitting scale in simulating future change in extreme precipitation (Under review) Vandentorren S, Bretin P, Zeghnoun A, Mandereau-Bruno L, Croisier A, Cochet C, Ribéron J, Siberan I, Declercq B, Ledrans M (2006) August 2003 heat wave in France: risk factors for death of elderly people living at home. Eur J Public Health 16:583–591. https://doi.org/10. 1093/eurpub/ckl063 Watkins R, Levermore GJ, Parkinson JB (2013) The design reference year—a new approach to testing a building in more extreme weather using UKCP09 projections. Build Serv Eng Res Technol 34:165–176. https://doi.org/10.1177/0143624411431170 Wilby RL, Wigley TML (1997) Downscaling general circulation model output: a review of methods and limitations. Prog Phys Geog 21:530–548. https://doi.org/10.1177/ 030913339702100403 Wilcox S, Marion W (2008) Users manual for TMY3. Renew Energy. doi:NREL/TP-581-43156 de Wilde P, Tian W (2010) Predicting the performance of an office under climate change: a study of metrics, sensitivity and zonal resolution. Energy Build 42:1674–1684. https://doi.org/10. 1016/j.enbuild.2010.04.011 Wilks DS, Wilby RL (1999) The weather generation game: a review of stochastic weather models. Prog Phys Geogr 23:329–357. https://doi.org/10.1191/030913399666525256 Wouters H, De Ridder K, Demuzere M, Lauwaet D, Van Lipzig NPM (2013) The diurnal evolution of the urban heat island of Paris: a model-based case study during Summer 2006. Atmos Chem Phys 13:8525–8541. https://doi.org/10.5194/acp-13-8525-2013 Wouters H, De Ridder K, Poelmans L (2016) Heat stress increase towards the mid-21st century twice as large for cities compared to rural areas 138 D. Ramon et al.

Wouters H, Demuzere M, Blahak U, Fortuniak K, Maiheu B, Camps J, Tielemans D, Van Lipzig NPM (2016) The efﬁcient urban canopy dependency parametrization (SURY) v1.0 for atmospheric modelling: description and application with the COSMO-CLM model for a Belgian summer. Geosci Model Dev 9:3027–3054. https://doi.org/10.5194/gmd-9-3027-2016 Wouters H, De Ridder K, Poelmans L, Willems P, Brouwers J, Hosseinzadehtalaei P, Tabari H, Broucke SV, van Lipzig NPM, Demuzere M (2017) Heat stress increase under climate change twice as large in cities as in rural areas: a study for a densely populated midlatitude maritime region. Geophys Res Lett 44:1–11. https://doi.org/10.1002/2017gl074889 Chapter 7 Evaluation of a Simpliﬁed Calculation Approach for Final Heating Energy Use in Non-residential Buildings

Barbara Wauman, Wout Parys, Hilde Breesch and Dirk Saelens

Abstract In Flanders, an obligatory software tool (EPR) is used to assess the energy performance of new buildings offering a simplified procedure to estimate the energy use for heating. This calculation approach is based on the principle of multiplying the building’s heating demand with standardised (sub)system efficiencies. In this paper, the accuracy of this simplified approach is assessed for a traditional, hydronic heating system in non-residential buildings. To do so, integrated dynamic simulations are performed in TRNSYS for a series of building design variants with varying insulation quality, thermal capacity, window-to-wall ratio and orientation. From the integrated simulations, monthly subsystem efficiencies are deduced. Results show that the efficiencies are significantly influenced by the part load ratio. As however losses of efficiencies are noticed only in periods of low heat demands, the overall effect on the annual use is limited. Energy assessment by the simplified method is within an error of <2.5 kWh/(m2Áa) or <10%. Therefore, the simplified approach as currently applied in the EPR calculation tool in Flanders is concluded to be suited for the calculation of the final energy use. An evaluation of tabulated values for the overall system efficiencies used in this simplified method is however recommended.

Keywords Energy use assessment for heating Á Integrated building and HVAC system simulations Á Non-residential buildings

B. Wauman Á H. Breesch (&) Department of Civil Engineering, KU Leuven, Construction Technology Cluster, Sustainable Building, Technology Campus Ghent, Gebroeders De Smetstraat 1, 9000 Ghent, Belgium e-mail: [email protected] W. Parys Bauphi, Stropkaai 55, 9000 Ghent, Belgium W. Parys Á D. Saelens Department of Civil Engineering, Building Physics Section, KU Leuven, Kasteelpark Arenberg 40 - Box 2447, 3001 Heverlee, Belgium D. Saelens EnergyVille, Thor Park 8310, 3600 Genk, Belgium

© Springer Nature Singapore Pte Ltd. 2019 139 E. Motoasca et al. (eds.), Energy Sustainability in Built and Urban Environments, Energy, Environment, and Sustainability, https://doi.org/10.1007/978-981-13-3284-5_7 140 B. Wauman et al.

7.1 Introduction

In the recently approved second recast of the Energy Performance of Buildings Directive (EPBD 2018), the European Union commits itself to develop a sustainable, competitive, secure and decarbonised energy system by 2050. An important aspect to achieve this goal is to find a cost-efficient equilibrium between decar- bonising energy supplies and reducing the final energy consumption of the building stock. This chapter focusses on the final energy use of non-residential buildings. To significantly reduce this energy consumption by 2050, the application of a reliable and accurate method for the energy use assessment of building designs is essential. In Flanders, an obligatory software tool (EPR) is used to assess the energy performance of new buildings. The energy performance of a building is determined based on calculated energy use reflecting typical energy use for heating, cooling, domestic hot water, ventilation and built-in lighting and other technical systems. The EPR-tool consists of two main parts: one aimed at residential buildings (VEA 2017) and one aimed at school and office buildings (VEA 2015). Both parts use the same calculation method and determine the energy performance of a building design in three consecutive steps: (i) the calculation of the net energy demand for heating and cooling, (ii) the calculation of the delivered energy to the heating and cooling systems as the division of the energy demand by the annually averaged, tabulated subsystem efficiencies for generation, storage, distribution, control and emission and (iii) the calculation of the primary energy use by adding the auxiliary energy needed for all system components and converting it to primary energy, taking into account renewable energy sources and national conversion factors which represent the conversion step from energy source to energy carrier. The energy balances are hereby calculated in steady-state conditions on a monthly time base. Particularly in the context of building energy regulations, the simplicity of the input, the transparency of the calculation rules, the intuitive correlation between input and output, the robustness and reproducibility of the quasi-steady-state method are considered as a great advantage compared to the more complex dynamic simulation tools (Van Dijk et al. 2005; Van Dijk and Spiekman 2004). On the other hand, dynamic effects are taken into consideration in a simplified way by time-weighted averaged values and empirically determined correction factors. Moreover, a simplified calculation approach is used to determine the final energy use. Each of the HVAC subsystems is evaluated separately and the subsystems’ thermal losses are calculated in a simplified way using tabulated efficiencies. These tabulated values are based on the related HVAC system’s components but are generally irrespective of the building characteristics and neglect the mutual inter- play between the various subsystems. The EPR-tool attempts to predict the buildings’ energy use as accurately as possible. However, the inherent model assumptions and simplifications might (over)simplify the dynamic and nonlinear interaction of the building and HVAC systems which may jeopardise the accuracy and the use of the method in the context of energy regulation. After all, to ensure the objectives of the building energy policy 7 Evaluation of a Simplified Calculation Approach … 141 and guarantee the EPB Directives effectiveness, the quasi-steady-state calculation results should be accurate and reliable. Moreover, as the EPR-tool is used for the execution of cost-optimal studies (Van der Veken et al. 2013; De Deygere and Troch 2013) in the context of EPBD and hence is used as a supporting tool for energy policy decision making, the calculation results should be realistic to avoid inaccurate cost-optimal design solutions and inefficient energy saving measures which could affect in turn the evolutions and trends on the building market (Pernigotto 2013). As argued by Zhang et al. (2006) and Van der Veken and Hens (2008), an integrated approach is therefore better suited, where a dynamic simulation is set up that includes both the building and the system. Though holistic studies on HVAC system performance characterisation are rather rare (Shahrestani 2013), this approach has been successfully adopted in a few studies, illustrating the influence of the building and the building use on the HVAC system efficiency. Korolija (2011) simulated a large amount of office buildings varying in orientation, insulation, glazing-to-wall ratio, glazing type, structural shading and daylighting coupled to four types of secondary HVAC systems for several weather data files using the integrated approach. The results indicated clearly that the annual heating and cooling system efficiency and the annual auxiliary energy use for the different systems are not constant but depend instead on the (thermal) characteristics of the building and on the building loads. Similarly, Van der Veken and Hens (2008), Peeters et al. (2008) and Van der Veken et al. (2006) demonstrated the relation between the insulation quality of the building envelope and the emission and distribution efficiency of heating systems in residential buildings. A rather strong relationship between the monthly total efficiency and the heat-balance ratio of the building—the ratio of heat gains overheat losses—was found. This effect is attributed to component efficiencies decreasing for lower part load ratios and overheating due to imperfect control occurring mainly when heating demand is low. Bauer (1999) established correlations between the heat emission and control efficiency and a parameter characterising the building properties for different combi- nations of systems. His findings were integrated in the German standard (DIN V 4701-10 2003). In line with these studies, this paper uses an integrated approach to evaluate the accuracy of the EPR-tool, focusing on schools and office buildings in particular (VEA 2015). To do so, the results from two similar but separate, larger studies— one on school buildings (Wauman 2015) and one on office buildings (Parys 2013) are brought together. Moreover, this paper focuses on the final energy use for heating. After all, in Flemish schools, heating is the most important contributor to the total energy use whereas active cooling1 and (de)humidification are rare (Wauman 2015). At the same time, a study from 2001 (BBRI 2000) which analyses in detail 50 recently renovated and constructed offices spread out over Flanders and

1Flanders has a temperate maritime climate influenced by the North Sea and the Atlantic Ocean with relatively moderate summers and mild winters. Moreover, schools are closed during July and August. 142 B. Wauman et al. hence demonstrates the market penetration at that time shows that 20% of office buildings are built without an active cooling system. To assess the accuracy of the simplified calculation method used in the EPR-tool, integrated building and HVAC system simulations of a relatively simple heating system commonly found in office and school buildings in Flanders are performed in TRNSYS. In extension to the research carried out by Korolija (2013), both the primary and secondary HVAC systems are modelled. The aim of this paper is to analyse the monthly subsystem efficiencies calculated with this integrated approach and to express them depending on building characteristics, in analogy with the work of Van der Veken and Hens (2008) and Peeters et al. (2008) for residential buildings, to acknowledge the influence of the building and building use on the HVAC system performance. The influence of these subsystem efficiencies on the EPR calculation results is then assessed to evaluate the accuracy of the simplified calculation approach.

7.2 Terminology

7.2.1 Conceptual Scheme

A heating system consists generally of four subsystems: a generation, storage, distribution and emission subsystem. The first two subsystems form the primary HVAC system, while the latter two form the secondary HVAC system. To calculate the final energy use for heating, the EPR calculation tool (VEA 2015) applies a sequential subsystem calculation approach (see Fig. 7.1a). Thermal losses are calculated separately for each of the included subsystems using tabulated subsystem efficiencies. The recoverable part of the thermal losses is hereby directly subtracted from the loss of each system and are thus accounted for by an increase of the related subsystems’ efficiencies (see Fig. 7.1b). This method is intuitive and

(a) Sankey diagram heating energy flows (b) Sub-system energy flow (Tchervilov and Kaloyanov 2012)

Fig. 7.1 Conceptual scheme of the simplified calculation procedure as currently applied in the EPR calculation tool (VEA 2015) 7 Evaluation of a Simplified Calculation Approach … 143 simple as no iterations are needed to simulate the performance of the building and its system. According to EPR, the final energy use for heating QH,final,use is calculated in two steps. First, the gross energy demand QH,gross is calculated which is the energy needed to be delivered by the generation system or plant to the secondary HVAC system (see Fig. 7.1a). Next, the total energy delivered to the heating system QH,final,use is calculated by dividing the gross energy demand QH,gross by the generation efficiency ηgen and the storage efficiency ηstor.

7.2.2 Calculation of the Gross Energy Demand

The calculation of the gross energy demand QH,gross depends on the yearly averaged efﬁciency of the secondary HVAC system ηsystem which covers both the waste of energy that occurs when a building is simultaneously heated and cooled, and the overall occurring thermal energy losses of the secondary HVAC system. General system efﬁciencies are calculated using Eq. 7.1:

1 nsystem ¼ ð7:1Þ 1 þ aheat þ fannih=fheat;net where aheat is a factor representing the heat losses due to distribution of heat and imperfect control of the heating system, fannih is a factor accounting for the amount of wasted energy due to simultaneous heating and cooling, fheat,net is the ratio of QH, nd and the sum of QH,nd and QC,nd. An overview of all standard tabulated values for the aforementioned factors based on the type of the heat distribution medium and the applied control system can be found in the EPR calculation manual Tables 9 and 10 (VEA 2015).

7.2.3 Calculation of the Final Energy Demand

The total energy input to the heat generation system QH,ﬁnal,use needed for the requested heating of the building is calculated using Eq. 7.2.

¼ QH;gross ð : Þ QH;final;use : 7 2 ggen gstor

The source of generation efficiency losses depends highly on the type of generation device. For boilers, for example, efficiency losses are caused by heat transfer through the chimney (or flue gas exhaust) and through the boiler’s wall. The latter depend in turn on the location of the boiler, the part load ratio of the heating system, the operational conditions of the boiler and the applied control strategy. In EPR, the yearly averaged generation efficiency ηgen is calculated based on the 30% part load 144 B. Wauman et al. ratio of the heating system and corrected for the design return water temperature using Eq. 7.3: ÀÁÀÁ ¼ þ : Á À ð : Þ ggen fHi=Hs g30% 0 003 h30% hmean;boiler 7 3 with

hmean;boiler ¼ 6:4 þ 0:63 Á hreturn;design ð7:4Þ where fHi/Hs is the net-to-gross conversion factor (e.g. fHi/Hs = 0.9 for natural gas), η30%,boiler is the 30% part load boiler efficiency based on the manufacturer’s data, h30% is the boiler supply flow temperature at which η30%,boiler is determined (°C) and hreturn,design is the design return flow temperature of the heating system (°C). The storage system efficiency ηstor is assumed to be equal to 1.

7.3 Building and HVAC System

Section 7.3.1 describes the typical building configuration and the users’ profile of a representative school and office building in Flanders. In Sect. 7.3.2, an overview of the properties of the selected HVAC system configuration and related control strategy are given. Both the characteristics as well as the specific details on the design and sizing procedure of the investigated HVAC system (components) are summarised.

7.3.1 Building Description

7.3.1.1 School Building Model

A representative (elementary) school building, based on statistical data gathered in previous research (Wauman et al. 2015), is used for this study. The school building is rectangular and consists of two floors of 1028.4 m2 useable floor area each, the floor plan of which can be found in Fig. 7.2. The floor-to-floor height is 2.8 m. The building’s total external height is 6.6 m. This results in a protected gross volume of 8147 m3, 3694 m2 loss surface (ground floor, roof and external walls) and a compactness level of 2.2 m. In Flanders, typically, lessons are evenly spread over five weekdays from Monday to Friday from days 8.30 a.m. till 4 p.m. Wednesday afternoon is free. Other activity and operational characteristics, typical for elementary schools, are 7 Evaluation of a Simplified Calculation Approach … 145

Fig. 7.2 Floor plan of a representative school building, partitioned in various heating (marked in grayscale) zones with external dimensions. Zonal heating set-point temperatures, set equal to the average heating temperatures in schools (Wauman et al. 2015), are marked summarised in Table 7.1. Moreover, the zonal set-point temperatures for heating are marked in Fig. 7.2. Additional information on the determination of the building’s characteristics and the deﬁnition of the boundary conditions can be found in Wauman et al. (2015).

7.3.1.2 Ofﬁce Building Model

A generic reference office building with cellular office spaces is assembled for this research, based on statistical data (BBRI 2000). The reference building is a detached office building with 4 floors of 500 m2 each, and the floor plan of which can be found in Fig. 7.3. The main axis of the building lies in east–west direction (the office zones facades facing south and north). The floor-to-floor height is 3.5 m, hence the building’s total height is 14 m. This results in a protected volume of 7000 m3, 2680 m2 loss surface and a compactness of 2.6 m. The imposed internal boundary conditions are described in Table 7.2. The zonal set-point temperatures for heating are marked in Fig. 7.3. 4 .Wua tal. et Wauman B. 146

Table 7.1 Overview of the boundary conditions typical for schools Occupancy Occupancy Ventilation rate IHG IHG Installed Partial Partial density persons appliances lighting power operation operation (m2/pers) (W/pers) (W/m2) (W/m2) appliancesa lightinga (%) (%) Classroom 8.30 a.m.–4 p.m. 329m3/(h pers) (S) 60 5 10.6 15 90 87.5% nominal capacity Ofﬁce 8.30 a.m.–4 p.m. 14 29 m3/(h pers) (S) 80 10 10 50 70 70% nominal capacity Canteen 12 a.m.–1 p.m. 1.5 29 m3/(h pers) (S) 60 2 6 100 95 Kitchen 8.30 a.m.–2.30 p.m. 10 36 m3/(h pers) (S)b 100 80 10 100 95 gym 8.30 a.m.–4 p.m. 20 44 m3/(h pers) (S) 160 – 6 – 95 Sanitary 8.30 a.m.–4 p.m. – 15 m3/(h m2) (E) –– 4 – 50 circulation 8.30 a.m.–4 p.m. – (E) –– 2 – 50 aAppliances and lighting are assumed to be switched on whenever a zone is occupied at a fraction of the nominal capacity, which are expressed by the partial operation factors b Extra ventilation due to heat producing cooking activities = 80 m³/(m²hood h) during 50% of the time 7 Evaluation of a Simpliﬁed Calculation Approach … 147

Fig. 7.3 Floor plan of a representative medium-sized ofﬁce building. Zonal heating set-point temperatures, set equal to the operative temperatures for a metabolic rate of 1.2 met and clothing value 1 according to (ISO 7730 2005), are marked

Table 7.2 Overview of the boundary conditions typical for office buildings Occupancy Occupancy Ventilation IHGc IHGc Installed densitya rateb occupied unoccupied lighting (m²/pers) (W/m²) (W/m²) powerd (W/m²) Offices 9 a.m.–9 p.m. 10.8 36 m³/(h pers) (S) 8.7 2 11 70% nominal capacity Meeting 3rd floor: 3.3 36 m³/(h pers) (S) 15 – 11 room 10 a.m.–11 a.m., 2 p.m.–3.30 p.m. Sanitary 8 a.m.–6 p.m. – 15 m³/(h m²) (E) –– 4.5 Storage ––3m³/(h m²) (E) –– 3.5 Circulation 8 a.m.–6 p.m. – (E) 8 1 3.5 aSensible heat gains 1 person = 75 W, 60% of which is convective. Latent heat gains 1 person = 55 W ASHRAE (2009) b(S) = supply, (E) = extraction. 95% of the supplied fresh air in a zone is assumed to be extracted cHeat gains due to appliances, based on Wilkins and Hosni (2000). Convective and radiative gains are split 50/50 dLights are assumed to be switched on whenever the zone is occupied. The internal gains are 50% convective and 50% radiative

7.3.2 HVAC System

In schools and ofﬁce buildings, a large variety of HVAC systems can be found. As system sizing and integrated simulations are very time-consuming, it is unfeasible to include all in this paper. The system selection is therefore limited to a hydronic heating system with high-temperature radiators, a natural supply and mechanical exhaust ventilation system and passive (night) cooling strategies2 to guarantee good

2An exhaust ventilation system, used for hygienic ventilation and passive cooling, is commonly found in contemporary Flemish schools (Wauman 2015) though is less typical for ofﬁce buildings. Nevertheless, as the study focuses on heating systems (see Sect. 7.1) and a hydronic heating system with radiators is regularly found in ofﬁce buildings, the same HVAC system is selected for both building typologies. 148 B. Wauman et al.

Fig. 7.4 Conceptual scheme of the investigated heating system configuration as it is implemented in TRNSYS for the school and office building. Real system configurations might deviate summer comfort. Lighting and their according control systems are only considered as a boundary condition (i.e. as a fixed part of the internal heat gains). Finally, as the need for hot water in schools and offices is generally limited, (centralised) production of domestic hot water is not considered. Separate heating circuits are foreseen for each of the building zones characterised by similar (time) trends of thermal loads and/or occupancy patterns. For the school building, for example, one heating circuit is foreseen for the administration zone, one for the canteen/kitchen, one for the gym and two heating circuits for the class zone as this zone is additionally split based on the orientation (see Fig. 7.2). The schematic, conceptual diagram as shown Fig. 7.4 presents in a simplified way the main components of the heating system. It comprises a modulating, condensing gas boiler and radiators controlled by thermostatic valves in every occupied zone. The heat distribution system consists of one central heating circuit and all secondary circuits coupled to a distribution header. The central and secondary circuits are decoupled using an open header. The radiators are sized according to the technical reports of the Belgian Building Research Institute (Debruyne 2013) and the European design heat load calculation standard (NBN 12831 2003). The typical characteristics of the selected radiator type are shown in Table 7.3. All distribution pipes are insulated according to the energy efficiency requirements for technical systems as described in VEA (2011). Pipe diameters are calculated allowing maximum friction losses of 100 Pa/m according to EN 15316-2-3 7 Evaluation of a Simplified Calculation Approach … 149

Table 7.3 Characteristics of Radiator characteristics Value the selected radiator type Radiator exponent, n 1.3 Height (m) 0.9 Specific nominal thermal power (75/65/20 °C) (W) 1961 Specific water content (l/m) 11.3 Specific weight (kg/m) 49.9 Nominal radiative fraction (%) 15

(2007). For the secondary circuits where the supply flow rate to the radiators is controlled by thermostatic radiator valves, variable speed pumps are installed. All pumps are sized based on both the design flow and the differential pressure for each of the served zones. The boiler heat output capacity is calculated as the sum of all nominal powers of the emission systems of each zone. A design outdoor temperature of −8 °C, representative for the Belgian climate, is applied and a safety margin of 10% for the design heat load capacity of the boiler is foreseen to cover for the distribution losses (NBN D 30-001 Centrale verwarming 1991). Heating is scheduled according to the occupancy profile. After school/working hours and during weekends, the heating is switched off. To reassure good thermal comfort at the start of the day, the heating starts at 7 a.m. each working day in the office building, and at 4 a.m. on Mondays and at 5 a.m. on the other school days in the school building. The zones without a set-point temperature marked (e.g. circulation) are not heated (see Figs. 7.2 and 7.3). An outdoor temperature reset control is applied on each of the secondary heating circuit separately. The maximum hot water supply temperature is 60 °C for the office building and 80 °C for the school3 Hence, high design heat loads are found in the school (max. 110 W/m2 in the school compared to max. 60 W/m2 in the office), which explains the selection of a higher maximum supply flow temperature. The set-point temperature of the boiler is then set equal to the maximum supply temperature of all the secondary heating circuits. An on/off burner control is applied based on the required supply flow set-point temperature: the boiler and boiler pump are turned on as soon as there is heat load and the boiler outlet temperature drops below a minimum (=boiler set-point −3 K). The boiler is switched off when there is no heat load or when the maximum boiler set-point temperature is reached (=boiler set-point +3 K). In between, the power is modulated according to the heat load pattern to reach the required set-point supply flow temperature. To avoid excessive cycling of the boiler, a minimum on/off time of 6 min is assumed. The pumps follow this boiler operation regime. A mechanical exhaust ventilation system is foreseen. Fresh outdoor air is supplied at a constant flow rate in the occupied rooms through trickle vents in the windows. The used air is extracted by extraction fans which are sized according to NBN 13779 (2010) (IDA 3). The operation of the ventilation fans is controlled by a

3Class rooms and canteens typically require high ventilation rates (see Table 7.4). 150 B. Wauman et al. time schedule according to the zonal occupancy proﬁles. For the school building, pre-ventilation prior to the school opening hours is applied to reassure good indoor air quality at any time of occupancy (EN 15251 2007). To maintain summer comfort, passive night cooling by mechanical ventilation is used, controlled by multiple control parameters: the night ventilation schedule 4 (0 h < time < 6 h), the star temperature as used in TRNSYS (hstar > hi;H;set, the outdoor air temperature, and the difference between operative and outdoor air temperature (hi;H;op − he > 2 °C) (Breesch 2006).

7.4 Method

To determine the actual energy performance of a heating system, a series of dynamic, integrated simulations is performed (see Sect. 7.4.1). Through these simulations, both the impact of the interaction between building and heating system (components) and the mutual interaction between the various subsystems are revealed. To assess the impact of the interaction between building and system, the investigated heating system is coupled to several building design variants. To obtain these variants, the building (envelope) properties are changed (see Sect. 7.4.2) while the overall design of the school and ofﬁce building remains the same. The energy performance of a heating system depends on the selection of heating system components and their characteristics, the applied sizing procedures, the heating settings and the control strategy used, and the quality of workmanship. For each of the selected systems, the appropriate sizing procedures are followed, the best market available components are selected and current good practice control decisions are made. Moreover, the systems are assumed to be well installed and commissioned. The implemented component models are all able to describe part load performance and the inﬂuence of non-rated conditions. Transient effects are included as much as possible, as the thermal inertia of the heaviest elements (boiler, radiator) and most of the water or air content of the system (pipes, ducts, boiler, radiator) are modelled. The system thermal losses both during and in between periods of operation are thus mostly accounted for in the simulations. In Sect. 7.4.3, the most important characteristics and input parameters used for the simulation of all HVAC system components (i.e. fans, pumps, pipes and boilers) are discussed.

4The star temperature, as used in TRNSYS (Klein et al. 2010), is the weighted average of the zone air temperature and the surface temperatures of the walls surrounding the zone. This star temperature differs from the operative temperature, which also is a weighted average of the air and mean radiant temperature, but with the weighing factor set equal to 0.5. 7 Evaluation of a Simpliﬁed Calculation Approach … 151

7.4.1 Integrated Dynamic Building and HVAC System Simulation

As TRNSYS 17 (Klein 2010) is well suited for HVAC system studies (Crawley 2005), it is selected as the dynamic simulation tool for this study. The Meteonorm weather file (Meteonorm 2005) of Brussels is used and the thermal behaviour is studied with a time step of 3 min for the school building, which is considered as a good balance between overall simulation time and precision (Peeters et al. 2008). For the office building, an even smaller time step of 1.5 min is used. Figure 7.5 illustrates in a simplified way the building and coupled heating system as it is implemented in TRNSYS. The considered data flow between the included system components is indicated by the arrows. The entire simulation model is then sequentially solved at every time step, iterating until convergence is reached. For each combination of building and system, two series of dynamic simulations are performed. Series 1 contains the dynamic simulations of the building envelope with ideal assumptions regarding the heating system operation (i.e. infinite heating power, perfect heating control, pure air heating) to calculate the ‘ideal’ energy demand QH;nd. To control the heating of a building zone, the operative temperature is maintained at the required set-point temperature level. As in the TRNSYS calculation procedure, the air temperature is used to control heating or cooling, a simple equation is implemented manipulating the air temperature so that the operative temperature fits the set-point requirements. Series 2 comprises the detailed simulations of both the building and the system and is executed to determine the ‘real’ energy use QH;final;use, incorporating the impact of the system (sizing), climatic conditions, and the implementation of operational schedules and control systems. In accordance with the simplified calculation approach (see Sect. 7.1), the results of the dynamic simulations are analysed at subsystem level (generation, distribution and emission) and are determined on a monthly time-base using Eqs. 7.2–7.6.

Fig. 7.5 Illustration of integrated building and HVAC simulation approach in TRNSYS 152 B. Wauman et al.

The gross energy demand depends on the heat demand of the building QH;nd and secondary HVAC system efficiency. The latter is determined by the combination of the emission and distribution losses. The emission losses are caused by temperature stratification, shielding of the emission devices, additional losses to the outside from heating devices embedded in the structure or due to imperfect control of the indoor temperature. The distribution losses are entirely due to thermal losses of the heat distribution pipes which depend on the thermal insulation and length of the distribution pipes, the average temperature of the heating medium and the temperature of the surrounding ambient. It is assumed that none of the pump power is converted into fluid thermal energy. The secondary HVAC system efficiency ηsys is calculated by dividing the energy demand (=results of dynamic simulations Series 1) by the gross heat demand (=result of dynamic simulations Series 2):

¼ QH;nd ð : Þ gsys 7 5 QH;gross

The energy delivered to the generation system or the final energy use for heating QH;final;use equals the ratio of the thermal energy output of the generation subsystem and the energy input by the energy carrier. The source of generation efficiency losses depends highly on the type of generation device. For boilers, for example, efficiency losses are caused by heat transfer through the chimney (or flue gas exhaust) and through the boiler’s wall. The latter depend in turn on the location of the boiler, the part load ratio of the heating system, the operational conditions of the boiler and the applied control strategy. The generation efficiency ηgen results from dividing the gross heating demand by the final energy use for heating or the heat input of the boiler.

¼ QH;gross ð : Þ ggen 7 6 QH;final;use

To link the HVAC system performance to the operation of the system and building properties, the calculated subsystem efﬁciencies of the considered HVAC system are expressed as a function of the part load ratio of the heating system b. This particular value is chosen as a reference as it incorporates the effect of the thermal insulation of the building, weather conditions and internal loads and as it is currently applied in EN 15251 (2007), EN 15316-2-1 (2007), EN 15316-4-1 (2008) to determine the subsystem efﬁciencies. The monthly averaged part load ratio of the heating system b is calculated according to Eq. 7.7:

¼ QH;nd ð : Þ b Ã 7 7 /boiler top where /boiler is the nominal power the boiler (kW) and top is the number of operational heating hours including the reheat period per month (h). In order to be 7 Evaluation of a Simpliﬁed Calculation Approach … 153 able to compare the performance of the different building design variants fairly, the resulting thermal comfort in the building needs to be similar in all cases. The thermal comfort is evaluated according to the degree hours’ criterion of EN 15251 (2007), with the difference between the occurring temperature and the limit temperature as weighting factor. Deviations during 5% of the occupied time on a yearly and monthly basis are accepted, which means about 60 h per year and about 7 h per month for the school buildings, and about 100 h per year and about 10 h per month for the ofﬁce buildings.

7.4.2 Building Model and Characteristics

Both investigated buildings are modelled as multi-zone buildings (TRNSYS Type56), divided based on the different users’ characteristics and heat load patterns of the included rooms. To avoid the excessive complexity of the simulation model and to limit the related calculation time, the number of included zones is limited: 1-person offices are modelled as one single zone per floor and orientation (see Fig. 7.3), classrooms are modelled as one zone per orientation (see Fig. 7.2). Moreover, only the third and fourth floor of the office building are modelled in the simulations, assuming their energy demands to be representative for the entire building. This simplification can be justified by the good thermal insulation values of both roof and floor. Both simplifications result in a multi-zone model with 14 thermal zones for the office building and a model with 6 thermal zones for the school building. To represent a wide range in annual net heating demand and assess the impact of the building’s characteristics on the heating (sub)system efficiencies, a selection of school and office building design variants is made (see Tables 7.4 and 7.5). Only those building variants are retained which guarantee good thermal comfort by passive cooling strategies.

7.4.3 HVAC System Model

An overview of the main (thermal) component models and related input parameters is given along this section. A detailed mathematical description and information on the input parameters of the models can be found in the TRNSYS manual (Klein 2010) and in the work of Parys (2013). The radiators are modelled by the dynamic, lumped capacitance radiator model TRNSYS Type362 (Holst 1996) that calculates both the emitted radiator power and the radiative fraction of the emitted power based on the water ﬂow, the surrounding temperature and the incoming water temperature. 5 .Wua tal. et Wauman B. 154

Table 7.4 Selection of 18 design variants of the school building the subsequent letters and numbers in the ﬁrst column refer to the thermal capacity of the building (Heavy–Light) the considered WWR (20 or 40%), the orientation (NS or EW) and the building energy efﬁciency level (variant 1–5)

Orientation Uwall Uroof Uglazing g-value n50 ACH WWR Shading Thermal (W/(m2K)) (W/(m2K)) (W/(m2K)) (−) (%) devicea capacityb H20NS_1 N-S 0.37 0.37 1.12 0.57 3 20 Fixed (S) Heavy H20NS_2 N-S 0.30 0.24 1.12 0.57 2.4 20 Fixed (S) Heavy H20NS_3 N-S 0.22 0.19 1.12 0.57 1 20 Fixed (S), mobile (E, W) Heavy H20NS_4 N-S 0.15 0.15 0.78 0.55 0.6 20 Fixed (S), mobile (E, W) Heavy H20NS_5 N-S 0.11 0.15 0.60 0.47 0.4 20 Mobile (S, E, W) Heavy H40NS_1 N-S 0.37 0.37 1.12 0.57 3 40 Fixed (S), mobile (E, W) Heavy H40NS_2 N-S 0.30 0.24 1.12 0.57 2.4 40 Fixed (S), mobile (E, W) Heavy H40NS_3 N-S 0.22 0.19 1.12 0.57 1 40 Fixed (S), mobile (E, W) Heavy H40NS_4 N-S 0.15 0.15 0.78 0.55 0.6 40 Fixed (S), mobile (E, W) Heavy H40NS_5 N-S 0.11 0.15 0.60 0.47 0.4 40 Mobile (S, E, W) Heavy L20NS_1 N-S 0.37 0.37 1.12 0.57 3 20 Fixed (S), mobile (E, W) Light L20NS_2 N-S 0.30 0.24 1.12 0.57 2.4 20 Fixed (S), mobile (E, W) Light L20NS_3 N-S 0.22 0.19 1.12 0.57 1 20 Fixed (S), mobile (E, W) Light H20EW_1 E-W 0.37 0.37 1.12 0.57 3 20 Fixed (S) Heavy H20EW_2 E-W 0.30 0.24 1.12 0.57 2.4 20 Fixed (S) Heavy H20EW_3 E-W 0.22 0.19 1.12 0.57 1 20 Fixed (S), mobile (E, W) Heavy H20EW _4 E-W 0.15 0.15 0.78 0.55 0.6 20 Fixed (S), mobile (E, W) Heavy H20EW_5 E-W 0.11 0.15 0.60 0.47 0.4 20 Mobile (S, E, W) Heavy aFixed = horizontally placed louvres. Mobile = automatically controlled screens, controlled on irradiation on façade: 250–150 W/m2 b Heavy = heavy walls, roof and floor; Cm = 95 Wh/(m²K). Light = heavy floor but light roof, external and internal walls; Cm = 43 Wh/(m²K) 7 Evaluation of a Simplified Calculation Approach … 155

Table 7.5 Selection of 5 design variants of the ofﬁce building varying the building envelope efﬁciency, shading and WWR. All main façades face north and south

Net heat Uwall Uroof Uglazing g-value n50 WWR Shading demand (W/(m²K)) (W/(m²K)) (W/(m²K)) (−) (ACH) (%) devicea (kWh/(m²a)) 22.6 0.30 0.30 1.1 0.59 2.5 21 No 20.1 0.20 0.20 1.1 0.26 1.0 31 No 26.5 0.40 0.30 1.1 0.44 2.5 21 Fixed (S) 33.7 0.40 0.30 1.1 0.26 2.5 71 Fixed (S) 37.0 0.60 0.40 1.1 0.29 2.5 31 Fixed (S) aFixed = horizontally placed louvres (TRNSYS type200a). Width slats = 20 cm, distance between slats = 19.5 cm

As in TRNSYS, single (zonal) air temperatures apply and only one-dimensional heat transfer are calculated, locally induced thermal losses of the heat emission systems or additional thermal losses due to temperature stratification in the room are not or hard to be simulated. Hence, to account for these losses, the emitted power of the radiators calculated in TRNSYS is adjusted before being coupled to the building model (TRNSYS Type56) by a fixed value, calculated according to EN 15316-2-1 (2007). One heat emitter with a total output capacity equal to the calculated design load is modelled per zone. To determine relevant distribution losses by the simulation model, however, a distribution pipe is coupled to the heat emitter with the length equal to the average of all distribution pipes needed to supply the real number of heat emitters. To determine the actual distribution losses, the thermal losses calculated for this pipe are then multiplied by the exact number of heat emitters that are foreseen in the considered building zone. The thermostatic radiator valves on the radiators are modelled based on the IEA annex 10’s perfect thermostatic valve model as developed by Ast (1986). It is a lumped capacitance model of the temperature sensor including the (relatively small) thermal resistance of the casing and the (larger) resistance between the sensor and the water. The valve authority is set equal to 0.7. The hysteresis is 0.5 °C and the nominal and maximal temperature differences between the valve and ambient are assumed to be 2 °C. All valves are assumed to have an infinite range ability: 100% of the maximum flow rate when the valve is fully open and 0% of flow rate when fully closed. For the simulation of the distribution pipes, the ‘plug-flow’ pipe model (TRNSYS Type31) is used. The surrounding temperatures of the pipes are simulated by the thermal simulation model and used as an input for the pipe models. In line with the simplified calculation approach as applied in EPR (VEA 2015), the distribution thermal losses which occur in heated zones are considered recoverable and are consequently considered injected as heat gains in the zones which they serve. 156 B. Wauman et al.

Table 7.6 Parameter values of gas boilers used in this research

Unom,max Unom,min Cburner VW UA (kJ/ Waux, Waux,max Waux,off (kW) (kW) (kJ/K) (m³) (K h)) min (W) (W) (W) Ofﬁce 142 47 94.5 0.221 46.52 45 185 30 School 311 104 128 0.279 77.22 55 385 30

For the dynamic simulations of the dynamic boiler, the boiler model developed by Haller et al. (2009) (TRNSYS Type869) is chosen. This model uses the incoming water flow rate and temperature as inputs and calculates the hot water outlet temperature and the related energy use. Parameters are the nominal range of thermal powers (/nom), the thermal capacitance of the burner (Cburner), the water content (Vw), the UA-value of the boiler and the auxiliary electrical energy use (W). The values for the boilers used in this research are listed in Table 7.6, taken from the technical data from a leading manufacturer. One should keep in mind that, despite the level of detailing of the integrated model, some realistic effects that influence the building energy use are neglected either due to inherent restrictions of the simulation model (e.g. realistic variations of the emission efficiencies cannot be calculated) or due to the modelling assumptions made (i.e. the impact of hydraulic imbalances or improper design on the efficiency of the HVAC system cannot be assessed in TRNSYS, so all components are assumed to be correctly installed and are connected to well-balanced hydraulic circuits which are however highly unlikely in real buildings).

7.5 Results and Discussion

In this section, the results of the integrated building and system simulations are discussed and analysed. For all cases, the desired thermal comfort levels (winter and summer) of all building zones are checked and found satisfactory, indicating that all system components were properly sized for every respective building design variant, and that the applied control strategy works well.

7.5.1 Final Energy Use for Heating

Figures 7.6 and 7.7 show the monthly final energy use for heating QH;final;use in relation to the monthly heating demand QH;nd for the school and office buildings respectively. Linear trend lines are added to both figures to assess the feasibility of the use of a constant value for the overall system efficiency to express the energy 7 Evaluation of a Simplified Calculation Approach … 157

Fig. 7.6 Monthly final energy use for heating (QH,final,use) expressed as a function of the monthly heat demand (QH,nd), both normalised to the building floor area (kWh/m2), for the 18 school building design variants

Fig. 7.7 Monthly final energy use for heating (QH,final,use) expressed as a function of the monthly heat demand (QH,nd), both normalised to the building floor area (kWh/m2), for the five office building design variants

use as a function of the heat demand (i.e. in conformity with EPR—see Sect. 7.1). The trend lines are forced to go through the origin to be physically correct. 2 Figures 7.6 and 7.7 show an almost linear correlation (R = 0.98) between QH; final;use and QH;nd, except for the lowest heat demands where the decrease of the system efficiencies causes a slight increase of the QH;final;use. As the decrease of efficiencies is mostly noticed in periods of low part load ratios and thus low heat demands, the effect on the annual energy use is limited. On the other hand, when comparing the results of the dynamic simulations with the results of the EPR-tool, significant differences are revealed (see Fig. 7.8). The results of the integrated dynamic simulations show that the thermal losses of the heating systems add approximately 25–30% to the net heating demand for school buildings. These additional losses are clearly underestimated by the 2 EPR-tool as QH,final,use,EPR is on average 16% or 11.8 kWh/(m a) lower than QH,final,use,dyn. The maximum difference of 29% is found for the light design variant with a WWR = 20% and the main axis oriented along the North–South direction (L20NS 1). Similar results are found for offices (Parys 2013). 158 B. Wauman et al.

Fig. 7.8 QH;ﬁnal;use expressed as a function of QH,nd, both normalised to the building ﬂoor area (kWh/(m2 a)), for HVAC1 coupled to 18 different school building design variants, calculated in TRNSYS using the EPR (VEA 2015) calculation standard

7.5.2 HVAC Subsystem Efﬁciencies

The results of the dynamically calculated and monthly averaged emission, distribution and generation efficiency for all school and office building design variants are shown in Fig. 7.9. To visualise the relation of the heating system performance to the operation of the system and the building properties, the efficiencies are plotted in relation to the part load ratio of the heating system b (see Eq. 7.7). Overall, three phenomena—similar for the school and office buildings—are noticed. (i) The efficiencies decrease significantly when part load ratios of the heating system b are lowered, especially noticeable for the emission efficiencies. Similar results were found earlier (Van der Veken et al. 2006; Bauer 1999) for emission efficiencies. Both studies indicate the decrease of control efficiencies at low part load ratios as the most important cause: when highly fluctuating internal and solar heat gains occur, accurate control of the heating system becomes more difficult so heat outputs result easily in overheating and affect negatively the control efficiencies. (ii) Figure 7.9a–f demonstrates that the emission efficiencies ηem are significantly influenced by the buildings’ characteristics whereas the distribution ηdis and generation efficiencies ηgen are not. To visualise the impact, two additional graphs are plotted which show the effect of the thermal capacity (heavy–light), the solar heat gains (i.e. WWR = 20–40%) and the insulation level (variant 1, 3 and 5—see Table 7.4) of the building on the emission efficiencies for the school building (see Fig. 7.10a, b). Figure 7.10a shows that overall slightly lower emission efficiencies are obtained in lighter buildings. Figure 7.10b shows that lower emission efficiencies are found for those school building variants with a window-to-wall ratio equal to 40%, hence for buildings with more solar heat gains. The impact of the insulation level of the building is not as straightforward as different results are found for the winter and the mid-season. In winter months, emission efficiencies are 7 Evaluation of a Simplified Calculation Approach … 159

Fig. 7.9 Monthly averaged ηem, ηdis and ηgen for all school and ofﬁce building design variants in function of the part load ratio of the heating system b

slightly higher in better insulated buildings. As however the heating system and related control settings are designed in accordance to the buildings’ characteristics (NBN EN 12831 2003), overall the impact of the insulation level remains limited. An opposite trend is noticed in spring and autumn months when higher solar heat gains occur and accurate control of the heating system becomes more challenging. As shown in Fig. 7.10b, this trend is more pronounced in more energy efﬁcient buildings. (iii) Overall, rather low values of the generation efﬁciencies ηgen are found. This can mostly be explained by the low part load ratios. According to the heating 160 B. Wauman et al.

Fig. 7.10 Impact of the buildings characteristics (i.e. thermal capacity, U-value and WWR) on the monthly averaged ηem in the school building

system design standard (NBN 12831 2003), boilers are sized as the sum of the design heat loads of all heated zones. In addition, a reheat capacity is included while, simultaneously, all solar and internal heat gains are neglected. Considering the typically high internal heat gains in classrooms and offices, this results in frequent low part load ratios, related excessive cycling and extra thermal losses of the boiler, which cause in turn a loss of the generation efficiency. Furthermore, the use of an open header in combination with a constant speed pump in the primary heating circuit negatively affects the generation efficiency: as the return water flow is mixed with the ‘unused’ part of the supply hot water flow, an increased return water flow temperature is found and hence lower efficiencies of the boiler are obtained. The thermal losses of the boiler related to the boiler operating conditions (i.e. expressed by the part load ratio b) and the return water flow temperature at bypass operation of the boiler are hence the important influencing factors and should be accurately modelled by the calculation method used for the assessment of the HVAC system performance. 7 Evaluation of a Simplified Calculation Approach … 161

The following overall conclusions can be drawn: • The subsystem efficiencies decrease significantly when part load ratios b are low. As the losses of the efficiency are only noticed in periods of low heat demands, it is shown that the overall effect on the annual energy use is be limited. • The control efficiency of the HVAC systems is affected by the characteristics of the building to which the HVAC system is coupled. Especially for the lighter buildings, a lower performance of the control efficiencies is noticed. Variations over the other three building design variants (i.e. H20NS, H40NS and H20EW) remain however limited to <10%. • The final energy use for heating is significantly underestimated by the EPR standard (VEA 2015) mainly due to a high overestimation of the generation efficiencies. Dynamic simulation results show that the generation efficiencies depend highly on the hot water temperature regime (i.e. heating curve, return temperature) and the part load ratios of the heating systems. Based on the comparative analysis of the static and dynamic calculated efficiencies, these effects appear however to be underestimated in the currently applied EPR calculation method. Adding supplementary data to take into account the specific boiler operation conditions of the individual installation and including the option to calculate the impact of recirculation (i.e. bypass) on the return water temperature could offer a better fit and could thus be a more accurate alternative for the EPR calculation standard. Other alternative heat generation, distribution and emission systems, likely to be found in contemporary offices and schools, are not studied here. Additional research is hence necessary to check if the use of the simplified calculation approach can be extrapolated to other HVAC system (configurations), control systems, etc. Moreover, the study focuses on the conceptual framework of the simplified calculation method only. Further research is necessary to assess the accuracy of the values of the yearly averaged subsystem efficiencies as currently defined in the simplified calculation standard.

7.6 Conclusion

This paper assesses the accuracy of the simplified calculation approach to determine the final energy use for heating in non-residential buildings in the framework of the EPBD. Integrated, dynamic simulations of a traditional HVAC system used in a school and office building are performed. The system consists of a modulating condensing gas boiler and radiators in every heated zone. No active cooling is provided; though acceptable summer comfort levels are guaranteed using passive (night) cooling strategies. Hygienic ventilation is foreseen by a mechanical exhaust ventilation system. To obtain a comprehensive set of analysis results, the 162 B. Wauman et al. investigated heating system is coupled to 18 school building design variants on the one hand, and 5 office building variants on the other hand. Based on the integrated simulation results, the influence of both the buildings’ characteristics and the system operation on the heating system performance is analysed. To do so, monthly efficiencies for the heat generation, distribution and emission subsystems are deduced and expressed as a function of the part load ratio of the heating system. Regarding the impact of the building’s characteristics, the results reveal that for the investigated, rather well-insulated building variants, the impact on the (sub)system efficiencies is limited. No impact is found on the heat generation and distribution efficiencies. The variations of the control efficiency of the heating emission systems due to varying characteristics of the building to which the system is coupled remain <10%. Regarding the heating system operation, a decrease of the (sub)system efficiencies is noticed when part load ratios of the heating system decrease. Nevertheless, as the losses of efficiency are only noticed in periods of low part load ratios and hence low heat demands, the overall effect on the annual final energy use for heating is limited. Finally, it is shown that the simplified calculation approach using a single coefficient to express the final energy use as a function of the heat demand offers good calculation results. For schools, the root mean square error is limited to 2 kWh/(m2 a) or 2.4%. For, the office buildings, the maximum RMSE is 2.4 kWh/(m2 a) or 9.9%. Despite the uncertainties and restrictions of the investigated simulation model, one may conclude the simplified approach as currently applied in the EPR calculation tool in Flanders is suited for the calculation of the final energy use. An evaluation of tabulated values for the overall system efficiencies used in this simplified method, based on more extensive dynamic simulation results, is however recommended.

References

ASHRAE (2009) ASHRAE fundamentals Ast H (1986) IEA—annex 10: thermostatic valves. Technical report. Stuttgart, Germany, Institut fur Kernenergetik un Energiesysteme, abt. Heizung, Luftung-Klimattechnik, Universitat Stuttgart Bauer M (1999) Methode zur Berechnung und Bewertung des Energieaufwandes fur die Nutzenubergabe bei Warmwasserheizanlagen. PhD thesis. Universitat Stuttgart, Germany BBRI (2001) Kantoor 2000—Studie van energiegebruik en binnenklimaat van kantoren (in Dutch). Technical report Breesch H (2006) Natural night ventilation in office buildings. PhD UGent, Belgium, Ghent, Belgium Crawley DB, Hand J, Kummert M, Griffith BT (2005) Contrasting the capabilities of building energy performance simulation programs (version 1.0). Technical report, July, US Department of Energy, University of Strathclyde Energy Systems Research Unit, University of Wisconsin Solar Energy Laboratory, National Renewable Energy Laboratory 7 Evaluation of a Simplified Calculation Approach … 163

De Deygere M, Troch E (2013) Studie naar kostenoptimale energieprestatie- eisen bij niet-residentiele gebouwen (in Dutch). Technical report. Brussel, Belgie, Vlaams Energie Agentschap (VEA) Debruyne R (2013) Rapport nr. 14: ontwerp en dimensionering van centrale-verwarmingsinstallaties met warm water (in Dutch). Technical report 14. Brussels, Belgium, WTCB DIN V 4701-10 (2003) Energy efficiency of heating and ventilation systems in buildings—part 10: heating, domestic hot water, ventilation. no. 0033 EN 15251 (2007) Indoor environmental input parameters for design and assessment of energy performance of buildings addressing air quality, thermal environment, lighting and acoustics EN 15316-2-3 (2007) Heating systems in buildings—method for calculation of system energy requirements and system efficiencies—part 2–3: space heating distribution systems EN 15316-2-1 (2007) Heating systems in buildings—method for calculation of system energy requirements and system efficiencies—part 2–1: space heating emission systems EN 15316-4-1 (2008) Heating systems in buildings. Method for calculation of system energy requirements and system efficiencies. Space heating generation systems, combustion systems (boilers) EPBD (2018) The revised energy performance of buildings directive (EU) 2018/844 Haller M, Konersmann L, Haberl R, Droscher A, Frank E (2009) Comparison of different approaches for the simulation of boilers using oil, gas, pellets or wood chips. In: 11th international IBPSA conference, Glasgow, Schotland, pp 732–739 Holst S (1996) TRNSYS—models for radiator heating systems, Munchen, Germany ISO 7730 (2005) Ergonomics of the thermal environment—analytical determination and interpretation of thermal comfort using calculation of the PMV and PPD indices and local thermal comfort criteria. http://linkinghub.elsevier.com/retrieve/pii/S0267726105000503 Klein SA, Beckham WA, Mitchell DW (2010) TRNSYS 17.1: a transient system simulation program. Solar Energy Laboratory, University of Wisconsin, Madison, USA Korolija I (2011) Heating, ventilating and air-conditioning system energy demand coupling with building loads for office buildings. PhD De Montfort University, Leicester, UK Korolija I, Marjanovic-Halburd L, Zhang Y, Hanby V (2013) UK office buildings archetypal model as methodological approach in development of regression models for predicting building energy consumption from heating and cooling demands. Energy Build 60:152–162. http://linkinghub.elsevier.com/retrieve/pii/S0378778812006810 Meteonorm ( 2005) Meteotest. Meteonorm versie 5.1 - Edition 2005 NBN EN 12831 (2003) Verwarmingssystemen in gebouwen - Methode voor de berekening van de ontwerp- warmtebelasting NBN EN 13779 (2010) Ventilation for non-residential buildings—performance requirements for ventilation and room-conditioning systems (in Dutch) NBN D 30-001 (1991) Centrale verwarming, ventilatie en luchtbehandeling - Gemeenschappelijke eisen voor alle systemen - Warmtegeneratoren en branders Parys W (2013) Cost optimization of cellular office buildings based on building energy simulation. PhD, KU Leuven, Belgium, Leuven, Belgium Peeters L, Van der Veken J, Hens H, Helsen L, D’haeseleer W (2008) Control of heating systems in residential buildings: current practice. Energy Build 40(8):1446–1455. http://linkinghub. elsevier.com/retrieve/pii/S0378778808000315 Pernigotto G (2013) Evaluation of building envelope energy performance through extensive simulation and parametrical analysis. PhD University of Padova, Italy Shahrestani M, Runming Y, Cook GK (2013) Characterising the energy performance of centralised HVAC&R systems in the UK. Energy Build 62:239–247. http://linkinghub. elsevier.com/retrieve/pii/S0378778813001849 Tchervilov L, Kaloyanov NG (2012) Study on energy efficiency in buildings in the contracting parties of the energy community—final report. Technical Report, February, Energy Saving International AS (ENSI) Van der Veken J, Hens H (2008) Determination of the heating efficiency at building level. In: building physics symposium, Liege, Belgium, pp 101–104 164 B. Wauman et al.

Van der Veken J, De Meulenaer V, Hens H (2006) Eindrapport GBOU - EL2EP PROJECT: Ontwikke- ling via levenscyclusoptimalisatie van extreem lage energie- en pollutiewoningen (in Dutch). Technical report Van der Veken J, Creylman J, Lenaerts T (2013) Studie naar kostenoptimale niveaus van de minimumeisen inzake energieprestaties van nieuwe residentiele gebouwen (in Dutch). Technical Report april. Geel, Belgium: Kenniscentrum Energie, Thomas More Kempen/KU Leuven Van Dijk HAL, Spiekman ME (2004) Energy performance of buildings: outline for harmonised EP procedures (ENPER - TEBUC). Technical report. Delft, The Netherlands, TNO Building and Construction Research Van Dijk HAL, Spiekman ME, De Wilde PA (2005) A monthly method for calculation energy performance in the context of European building regulations. In: 9th international IBPSA conference, Montreal, Canada, pp 255–262 VEA (2011) Bijlage XII: Systeemeisen (in Dutch). Brussels, Belgium VEA (2015) Annex VI Bepalingsmethode van het peil van primair energieverbruik van kantoor- en school- gebouwen (in Dutch). Brussels, Belgium VEA (2017) Annex V Bepalingsmethode van het peil van primair energieverbruik van woongebouwen. Brussels, Belgium Wauman B (2015) Evaluation of the quasi-steady-state method for the assessment of energy use in school buildings. PhD thesis, KU Leuven Wauman B, Saelens D, Breesch H (2015) The definition of representative boundary conditions for Flemish schools for use in energy assessment methods. Energy Build 87:1–13. http:// linkinghub.elsevier.com/retrieve/pii/S0378778814008780 Wilkins C, Hosni MH (2000) Heat gain from office equipment. ASHRAE J 42:33–44. New York Zhang Y, Wright JA, Hanby VI (2006) Energy aspects of HVAC system configurations—problem definition and test cases. HVAC R Res 12(3C):871–888 Chapter 8 A Review of Fuzzy-Based Residential Grid-Connected Microgrid Energy Management Strategies for Grid Power Profile Smoothing

Diego Arcos-Aviles , Francesc Guinjoan, Julio Pascual, Luis Marroyo, Pablo Sanchis, Rodolfo Gordillo, Paúl Ayala and Martin P. Marietta

Abstract Residential grid-connected microgrids (MG) comprising renewable generation and storing capability are constrained to grid-operator requirements which include, among others, a smooth and bounded grid power profile. These requirements attempt to mitigate a high unpredictability on the electrical power exchanged between the grid and the MG and affect the design of the MG Energy Management System (EMS). This chapter reviews several energy management strategies based on Fuzzy-Logic Controllers (FLC) designed in the last years to smooth the grid power profile of a residential grid-connected MG. Two MG power architectures are considered. Both include wind and PV solar renewable generation and non-controllable domestic electrical loads. The first architecture assumes a battery charger/inverter as the only controllable element whereas the second one also considers a thermal load as an additional controllable element. The chapter presents a fuzzy logic approach to design the control strategies of the microgrid EMS. The strategies are designed under two scenarios, the first one assuming that forecast of generation and consumption is not available and the second one using MG forecasted data. Simulation and experimental results are provided to highlight and compare the features of all the strategies in terms of their power profile smoothing capability.

D. Arcos-Aviles (&) Á R. Gordillo Á P. Ayala Departamento de Eléctrica y Electrónica, Universidad de las Fuerzas Armadas ESPE, Av. Gral. Rumiñahui s/n, Sangolquí 171-5-231B, Ecuador e-mail: [email protected] F. Guinjoan Á M. P. Marietta Department of Electronics Engineering, Escuela Técnica Superior de Ingenieros de Telecomunicación de Barcelona, Universitat Politècnica de Catalunya, C. Jordi Girona 31, 08034 Barcelona, Spain J. Pascual Á L. Marroyo Á P. Sanchis Department of Electrical and Electronics Engineering, Public University of Navarre (UPNa) Ediﬁcio de los Pinos, Campus Arrosadia s/n, 31006 Pamplona, Spain

© Springer Nature Singapore Pte Ltd. 2019 165 E. Motoasca et al. (eds.), Energy Sustainability in Built and Urban Environments, Energy, Environment, and Sustainability, https://doi.org/10.1007/978-981-13-3284-5_8 166 D. Arcos-Aviles et al.

Keywords Distributed power generation Á Renewable power Á Energy management Á Power forecasting Á Fuzzy logic control Á Microgrids Á Power smoothing

8.1 Introduction

Several governments are implementing different policies to encompass a sustainable energy future as a primordial societal goal. This is the case for instance of the European Union which has planned a “clean energy for all Europeans” strategy for the next years (European Commission 2018). For the case of electrical energy, these policies promote the increasing use of renewable energies and their integration into the existing electrical system network. This chapter reviews some strategies for grid-connected residential microgrid systems including Renewable Energy Sources (RES) which take care of the power profile exchanged with the existing mains. These strategies contribute to the energy sustainability in two ways. On the one hand they allow the customers renewable electricity self-consumption. On the other hand the control of the grid power profile will facilitate the acceptance of Distribution System Operators (DSO) concerning the grid connection of such systems and, consequently, their integration into the existing electrical network. Microgrids (MG) are becoming a promising alternative to traditional centralized electric systems both in technologically advanced countries and developing ones (Tuballa and Abundo 2016; Asmus et al. 2013; Schnitzer et al. 2014). Microgrids, as defined in Hatziargyriou (2014), Lasseter (2002), include loads, Distributed Generation (DG) systems (including renewable sources), Energy Storage System (ESS), and are managed as a single unit by an Energy Management System (EMS) in order to exchange power with the grid through a single Point of Common Coupling (PCC). In other words, a MG constitutes an electrical power architecture connecting sources, loads and storage elements where some of these units can be controlled by the EMS. Microgrids are applied in different types of scenarios and can be classified according to their operation as either grid-connected or stand-alone (i.e., islanded), or according to their use as military, industrial, residential, etc. Regarding to their DG elements, they will comprise either renewable generators (e.g., photovoltaic panels and wind turbines) or fossil fuel-based generators (e.g., diesel generating sets or gas micro-turbines). The diversity of microgrids encompasses a wide range of configurations and goals, which will determine the energy management strategy governing each particular microgrid (Pascual et al. 2015). This chapter considers a grid-connected residential microgrid which includes wind and photovoltaic renewable generation, a lead-acid battery-bank as storing element as well as electrical and thermal loads. The microgrid is connected to the low voltage grid distribution line through a bidirectional multiport power converter. There is a wide variety of works dealing with different EMS strategies and power architectures. A comprehensive review of some of the latest papers regarding microgrids’ energy management can be found in Olatomiwa et al. (2016). Some 8 A Review of Fuzzy-Based Residential Grid-Connected Microgrid … 167 papers focus on minimizing the overall operating cost based on the cost of operation of every unit as well as the market electricity price (Xue et al. 2014; Comodi et al. 2015; Velik and Nicolay 2014; Tascikaraoglu et al. 2014; Niknam et al. 2012; Chen et al. 2013; Parisio et al. 2014). However, market electricity prices depend on technical issues, mainly, the primary source for generating electricity which in part may depend on the saturation of the transmission system. For this reason, energy management strategies reducing generation and consumption peaks in the power profile exchanged with the grid (i.e., “smoothing” the grid-power profile) result of interest. These strategies can be achieved balancing energy generation and consumption locally by using energy storage and controllable loads. Moreover, grid power fluctuations are reduced by making use of the ESS. Under the assumption of a proper ESS operation, by focusing on this technical aspects, results in the following outcomes: (1) reduction of overvoltage events in low-voltage grids when power is injected into the main grid (Masters 2002; Mahmud et al. 2011); (2) saturation alleviation in transmission lines (Black and Larson 2007); and (3) reduction in power fluctuations which leads to better grid quality and stability (Parissis et al. 2011; Shinji et al. 2008). In addition, smoothing the grid power profile can facilitate the grid operators control and, consequently, the penetration of RES into the distribution network. Early works facing grid power profile smoothing are based in an open loop processing. These works proposed a Finite Impulse Response (FIR) Simple Moving Average filter (SMA) to process the load demand and renewable generation power balance of the MG which will be injected (absorbed) to (from) the mains (Kim et al. 2008; Zhou et al. 2011). This filter decouples the MG power balance frequency spectrum delivering the high frequency components to the ESS and the low frequency ones to the mains to achieve the grid power profile smoothing. The main shortcoming of this approach lies on the delay introduced by the SMA filter window size which is designed for a power balance time series data of one day length. Indeed, consecutive sunny days would result in a net power injection from the MG to the mains. However, if the next day is a cloudy one the MG would still deliver an amount of energy according to the previous sunny days since the SMA filter should wait 24-h to take into account the data corresponding to the cloudy environmental conditions. This delay entails the possibility of the ESS over discharge failure and, consequently, the grid power profile will correspond to that resulting from the MG power balance. In order to preserve the proper ESS operation, a second approach complements the previous one by inserting a control loop to control the State-of-Charge (SOC) of the ESS (Barricarte et al. 2011). The design of this loop includes simple analytical functions fixing the range of the SOC variation, being the parameters of these functions heuristically adjusted. However, this strategy pre- serves the ESS operation at expense of grid power high fluctuations. On the other hand, the work in Barricarte et al. (2011), was upgraded in Pascual et al. (2014) by integrating the thermal part of the microgrid in the energy management system. In this way the controller now manages a water heater to enhance the use of the battery taking advantage of the hot water tank ability to store energy. The grid power profile was satisfactorily smoothed by using such demand-side 168 D. Arcos-Aviles et al. management technique. This approach suggests the importance of integrating the thermal subsystem of a microgrid in the energy management system due to three main reasons: (1) it comprises controllable electrical high-power devices; (2) it usually handles a big fraction of the energy used in microgrids; and (3) it usually provides extra energy storage in the form of a hot water tank or other forms of heat or cold storage. As in Barricarte et al. (2011), the parameters involved in the design of this EMS strategy are heuristically adjusted. To further improve these previous strategies in Pascual et al. (2015) the moving average approach was brought back but, in order to avoid the lag introduced by the SMA in the grid power, the strategy was instead based on the Central Moving Average (CMA). Such strategy needs forecasted values for the power profiles in the microgrid, which are calculated using weather forecasting for renewable generation, and persistence techniques for demand profiles. In this case the EMS strategy takes into account the forecasting error for the battery SOC control to better smooth the grid power profile. As in the previous cases the best performance is achieved by heuristically adjusting the parameters of the EMS strategy. From the control point of view, the aforementioned strategies highlight that both the grid power profile and the battery state of charge are the variables to be concurrently controlled and consider different variables as a control ones (for instance, the generated net power, the battery SOC or the forecast error in Pascual et al. 2015). The design of the resulting control loops is based on a heuristic knowledge of the desired behavior of the MG and use heuristically adjustable analytical expressions. This heuristic knowledge suggests the use of Fuzzy Logic Controllers (FLC) for the design of the EMS strategy in a residential grid-connected scenario. In this regard, fuzzy logic provides a formal methodology for representing, manipulating, computing, and implementing a human’s heuristic knowledge about how to control a system (Passino and Yurkovich 1998; Serraji et al. 2015). In addition, it is a powerful control technique capable of dealing with the imprecisions and nonlinearity of complex systems, that can be based on experience of the user about the system behavior rather than the mathematical model of the system as in the traditional control theory (Arcos-Aviles et al. 2016b; Mohamed and Mohammed 2013; Fossati et al. 2015). Based on these previous works, this chapter presents the design of fuzzy-based EMS strategies for a grid-connected residential microgrid for two MG power architectures to improve the grid power profile keeping the storage element operation between secure limits. Both include wind and PV solar renewable generation and non-controllable domestic electrical loads. The first architecture assumes a battery charger/inverter as the only controllable element whereas the second one also considers a thermal load as an additional controllable element. Both approaches are designed under two scenarios, the first one assuming that forecast of generation and consumption is not available and the second one using forecasted data, this resulting in four different fuzzy-based EMS strategies. Collecting the results of (Arcos-Aviles et al. 2012, 2014a, b, 2015, 2016b, 2017a, b, 2018), the chapter addresses in an unified way the complete design of the main fuzzy controllers parameters for each strategy, namely: (a) input and output variables of the 8 A Review of Fuzzy-Based Residential Grid-Connected Microgrid … 169

FLC, (b) membership functions for the FLC inputs and outputs (number, type and mapping) and (c) FLC rule base. These designs always try to reduce the controller complexity, minimizing the resulting number of rules. Section 8.2 is devoted to the description of the microgrid under study in terms of architecture and power sizing. Section 8.3 describes the mathematical modelling of the different elements comprising in the analyzed power architectures. Section 8.4 presents a set of quality indicators to evaluate the smoothness level of the grid power proﬁle. Section 8.5 addresses the design of the four FLC-based EMS strategies. Section 8.6 presents the simulation results of the features of each strategy and compares the features of all the strategies in terms of both their power proﬁle smoothing capability and battery SOC evolution. Finally the last section draws the conclusions of this chapter.

8.2 Microgrid Power Architectures and Components Modeling

This chapter considers two microgrid power architectures namely electric and electro-thermal microgrid. Both include a Hybrid Renewable Energy System (HRES) which comprises a photovoltaic (PV) generator of 6 kWp and a small wind turbine (WT) of 6 kW, domestic load with a rated power of 7 kW, and ESS which includes a lead-acid battery bank with a rated capacity of 72 kWh. In the following, both architectures are described. In addition, this section also introduces the analytical models evaluating the renewable power generation, as well as the battery SOC and the thermal storage and water temperature estimators needed for the EMS strategies based on forecasted data.

8.2.1 Electric Microgrid

The ﬁrst power architecture under study consists on a residential grid-connected electric MG (Pascual et al. 2015; Arcos-Aviles et al. 2012, 2016b, 2017a, 2018) and is shown in Fig. 8.1. Note that the arrows direction depicted in Fig. 8.1 implies a positive power ﬂow. In this framework, the total renewable power generation, PGEN, the wind turbine power, PWT, the photovoltaic power, PPV, and the load power, PLOAD, are always positive. Conversely, the battery power, PBAT, is positive when the ESS delivers power to the MG (i.e., battery discharging process) and it is negative otherwise (i.e., battery charging process). Meanwhile, the grid power, PGRID, is positive when the grid delivers power to the MG and is negative when it absorbs power from it. 170 D. Arcos-Aviles et al.

The grid power proﬁle of the ﬁrst architecture can be expressed as the difference between the MG net power, PLG, and the battery power, as follows:

PGRID ¼ PLG À PBAT ð8:1Þ

PLG ¼ PLOAD À PGEN ð8:2Þ

PGEN ¼ PPV þ PWT ð8:3Þ

8.2.2 Electro-thermal Microgrid

The second power architecture consists on a residential grid-connected electro-thermal MG (Pascual et al. 2014; Arcos-Aviles et al. 2014a, 2017b). This architecture takes into account the same elements described in the first architecture and includes a Domestic Hot Water (DHW) system comprising of an Electric Water Heater (EWH) of 2 kW, a flat-plate collector of 2 kW, a thermal storage system represented by a hot water tank of 800 L capacity, and a thermal load (i.e., DWH consumption) equivalent to 2 kW, as shown in Fig. 8.2. From Fig. 8.2, on the DHW sub-system PH,e represents the power required by the EWH to keep the water temperature in the thermal storage between 45 and 65 °C, QWH is the rate of energy transferred from the EWH to the hot water tank, QSC is the rate of energy collected from the flat-plate collector, and QDHW is the domestic hot water consumption (Arcos-Aviles et al. 2017b). 8 A Review of Fuzzy-Based Residential Grid-Connected Microgrid … 171

Fig. 8.2 Second architecture under analysis, residential grid-connected electro-thermal microgrid

Similarly, the arrows direction in Fig. 8.2 denotes a positive power ﬂow. Note that the analysis of an electro-thermal MG implies the study of all thermal variables involved in the DWH system, so that, a complete analysis should contemplate the thermal losses in the hot water tank, namely QLOSS. The power exchanged with the grid of the electro-thermal MG, shown in Fig. 8.2, is expressed as follows:

PGRID ¼ PLOAD À PGEN À PBAT þ PH;e ð8:4Þ

8.3 Components Modeling

This section presents the mathematical modelling of all elements that are part of the aforementioned power architectures where Sects. 8.3.1, 8.3.2 and 8.3.3 are used in both architectures (Figs. 8.1 and 8.2) whereas Sects. 8.3.4 and 8.3.5 are used only in the second power architecture under study (Fig. 8.2). 172 D. Arcos-Aviles et al.

8.3.1 Photovoltaic Modelling

The output power of the PV array can be expressed as follows (Pascual et al. 2015; Arcos-Aviles et al. 2016b, 2017a; Fathima and Palanisamy 2015; Zhao et al. 2013; Yoo et al. 2012):

Gðb; aÞ PPV ¼ PSTC Á Á ½ð1 þ cðÞTC À TSTC 8:5Þ GSTC where PSTC is the PV module output power (W) under Standard Test Conditions 2 (STC), G(b, a) is the incident irradiance on the plane of the panels (W/m ), GSTC is 2 incident irradiance under STC (W/m ), c is the power temperature coefﬁcient, TSTC is the temperature under STC (°C), and TC is the cell temperature (°C), which is deﬁned considering the ambient temperature, Ta (K) and the Nominal Operating Cell Temperature (NOCT) (°C), as follows (Pascual et al. 2015; Arcos-Aviles et al. 2017a):

GðÞb; a T ¼ ðÞþT À 273 Á ðÞNOCT À 20 ð8:6Þ C a 800

Note that to compute the PV power by means of (8.5) and (8.6) it is necessary to transform the solar irradiance on a horizontal plane provided by the local obser- vatories, G0, into the solar irradiance on a tilted surface, G(b, a), following the procedure described in Lorenzo (2011).

8.3.2 Wind Turbine Modeling

The output power of a wind turbine is deﬁned as follows (Pascual et al. 2015; Arcos-Aviles et al. 2016b, 2017a; Manwell et al. 2009; Chong et al. 2014):

1 3 P ¼ Á q Á A Á C ; Á v ð8:7Þ WT 2 p WT ðZÞ

3 where q is the air density (kg/m ), v(Z) is the wind speed at the wind turbine 2 hub-height (m/s), A is the rotor swept area (m ), and CP,WT is the WT power coefﬁcient (Pascual et al. 2015; Arcos-Aviles et al. 2016b, 2017a). Note that environmental conditions, such as temperature, atmospheric pressure, among others, can affect the air density, so that, it should be estimated assuming the air behavior as an ideal gas, as follows (Pascual et al. 2015; Arcos-Aviles et al. 2016b, 2017a; Mathew 2006): 8 A Review of Fuzzy-Based Residential Grid-Connected Microgrid … 173

M Á p q ¼ ð8:8Þ Ta Á R where M is the air molar mass in kg/mol, p is the atmospheric pressure (N/m2), and R is the universal gas constant [J/(K mol)]. Note that to ﬁnd the estimation of the WT power is necessary to extrapolate the wind speed given by the local observa- tories (usually wind speed measured at 10 m height above the ground level) to the wind speed at the WT hub-height, as follows (Pascual et al. 2015; Arcos-Aviles et al. 2016b, 2017a; Mathew 2006; Ally et al. 2015):

ðÞ= ¼ Á ln Z Z0 ð : Þ vðZÞ vðZREF Þ 8 9 lnðÞZREF=Z0

being VðZREF Þ the wind speed at ZREF height, ZREF the height at the measured data, Z the turbine hub-height, and Z0 the roughness index of the terrain (Danish Wind Industry Association 2003).

8.3.3 Battery SOC Estimator

The battery SOC can be estimated as follows (Arcos-Aviles et al. 2012, 2014a, b, 2015, 2016b, 2018, 2017a; b; Tazvinga et al. 2015):

SOCðnÞ¼SOCðn À 1ÞÀDSOCðnÞð8:10Þ g DSOCðnÞ¼ Á PBAT ðn À 1ÞÁTs ð8:11Þ CBAT = ; 8 [ ¼ 1 gD PBAT 0 ð : Þ g ; 8 \ 8 12 gC PBAT 0 where the indices n and (n − 1) represent the current and the previous samples, respectively, DSOC refers to the battery SOC variation during the sampling period Ts, η is the battery efﬁciency (i.e., discharge and charge efﬁciencies, ηD and ηC, respectively), and CBAT is the battery rated capacity (Arcos-Aviles et al. 2017a, 2018). Note that the battery SOC must be kept between secure limits to preserve the ESS lifetime. Therefore, a maximum Depth of Discharge (DOD) of 50% is considered since the MG architectures described in this chapter includes a lead-acid battery bank (Guasch and Silvestre 2003; Anuphappharadorn et al. 2014). Accordingly, the battery constraints are expressed as follows: 174 D. Arcos-Aviles et al.

SOCMIN SOCðnÞSOCMAX ð8:13Þ

SOCMIN ¼ ðÞÁ1 À DOD SOCMAX ð8:14Þ being SOCMIN and SOCMAX the minimum and maximum battery SOC limits, respectively.

8.3.4 Flat-Plate Collector Modeling

The rate of useful energy collected from the ﬂat-plate collector, QSC, is expressed considering the technical characteristics sheet given by the manufacturer. For instance, the QSC of the ﬂat-plate collector used in the second power architecture is expressed as follows (Arcos-Aviles et al. 2014b, 2017b):

ðÞ; h G b a À9 6 À7 5 À5 4 QSC ¼ À2 Â 10 Á ðÞTWD À Ta þ 6 Â 10 Á ðÞTWD À Ta À7 Â 10 Á ðÞTWD À Ta þ ÁÁÁ GSTC i 3 2 ÁÁÁ þ0:003 Á ðÞTWD À Ta À0:07 Á ðÞTWD À Ta À8:6968 Á ðÞþTWD À Ta 1735:5 ð8:15Þ where TWD is the water temperature (°C). Note that the ambient temperature in (8.15) is given in °C.

8.3.5 Thermal Storage and Water Temperature Estimator

The storage capacity of the hot water tank, QST, is used to stablish the thermal energy balance in the MG, as follows (Arcos-Aviles et al. 2017b):

QST ðnÞ¼QSCðnÞþQWHðnÞÀQLOSSðnÞÀQDHW ðnÞð8:16Þ ð Þ¼ Á Á Á D ð Þð: Þ QST n qW CP V TWD n 8 17

2p Á L Á kT QLOSSðnÞ¼ Á ½ðTWDðnÞÀTEXT ðnÞ 8:18Þ lnðÞR1=R2 where L is the length of the hot water tank, kT is the hot water tank thermal conductivity, R1 and R2 are the inner and outer radius of the hot water tank, TEXT is the environment temperature where the thermal storage is located, qW is the density of water, CP is the speciﬁc heat capacity of water, V is the thermal storage volume, ΔTWD is the water temperature variation during a sampling period Ts, and TWD is the current temperature in the hot water tank after sampling period Ts (Arcos-Aviles et al. 2017b). 8 A Review of Fuzzy-Based Residential Grid-Connected Microgrid … 175

In this regard, assuming that the heat losses in the thermal storage are constant at each sampling step, the current temperature in the hot water tank can be estimated as follows:

ð Þ¼ ð À Þþ Ts Á ½ð Þþ ð ÞÀ ð ÞÀ ð Þ TWD n TWD n 1 Á Á QSC n QWH n QLOSS n QDHW n qW CP V ð8:19Þ

8.4 Grid Power Proﬁle Quality Criteria

The grid power profile quality criteria are defined to quantify the improvement of the power profile exchanged with the grid obtained by a specific energy management strategy. The quality criteria are defined in Pascual et al. (2015), Arcos-Aviles et al. (2017a, 2018) as follows:

8.4.1 Positive and Negative Grid Power Peaks

They are deﬁned as (Arcos-Aviles et al. 2017a):

PG;MAX ¼ maxðÞPGRID ð8:20Þ

PG;MIN ¼ minðÞPGRID ð8:21Þ where PG,MAX and PG,MIN are the maximum power delivered by the grid and the maximum power injected to the grid measured during the year, respectively. They are expressed in kW.

8.4.2 Maximum Power Derivative (MPD)

It is the maximum grid power proﬁle ramp-rate (i.e., the slope in two consecutive samples, being the sampling period Ts = 900 s) in the year under study (Arcos-Aviles et al. 2017a). The MPD criterion is expressed in W/h and it is computed as: ÀÁ _ MPD ¼ max PGRID ð8:22Þ 176 D. Arcos-Aviles et al. where ṖGRID is the grid power proﬁle ramp-rate, i.e.: ð ÞÀ ð À Þ _ PGRID n PGRID n 1 PGRID ¼ ð8:23Þ Ts

8.4.3 Average Power Derivative (APD)

It is the annual average value, in absolute value, of the grid power proﬁle ramp-rates (Arcos-Aviles et al. 2017a). The APD criterion is expressed W/h, as follows:

XN 1 _ APD ¼ PGRIDðnÞ ð8:24Þ N n¼1 being N the number of samples in one year.

8.4.4 Power Proﬁle Variability (PPV)

This criterion measures the grid power profile steadiness. It is calculated as follows: qPffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ff 2 f ¼f PGRID;f PPV ¼ i ð8:25Þ PDC where PGRID,f is grid power harmonic at f frequency, fi and ff are the initial and final frequencies, respectively, and PDC is the yearly power average value. This indicator −6 only evaluates frequencies above fi = 1.65 Â 10 Hz (i.e., one week or less variation periods), since the energy management strategies described in this chapter seek to compensate daily variations. The maximum frequency used to calculate −4 PPV is half of the sampling frequency ff = 5.55 Â 10 Hz (i.e., Nyquist frequency) (Pascual et al. 2015; Arcos-Aviles et al. 2017a, 2018).

8.5 Fuzzy-Based Energy Management Strategies

8.5.1 EMS Based on MG Energy Rate-of-Change (ERoC)

This strategy was presented in Arcos-Aviles et al. (2015) and an extended version including experimental validation results was described in Arcos-Aviles et al. (2018). The goal of this strategy, as well as the goal of the other strategies that will be further described in this chapter, is to smooth the grid power profile while keep 8 A Review of Fuzzy-Based Residential Grid-Connected Microgrid … 177 the ESS operation within secure limits, which is done by including a FLC within the energy management strategy. In this regard, the key factor of this strategy is the use of the MG Energy rate-of-change (ERoC) as a new input of the FLC. The ERoC allows quantifying the magnitude of the MG energy changes by means of measuring the slopes produced in two consecutive samples of the MG average power profile. It is worth noting that the ERoC can be understood as the local prediction of the battery SOC future behavior if the grid power is not modified (Arcos-Aviles et al. 2018). Accordingly, the FLC uses the ERoC value and the battery SOC to increase, decrease or maintain the power delivered or absorbed by the mains to concurrently fulfill the load power demand and to satisfy the constraints imposed by the ESS. Control strategy. Considering the first power architecture depicted in Fig. 8.1, the control strategy of the grid power profile is defined as the sum of two components (Arcos-Aviles et al. 2018):

PGRIDðnÞ¼PAVGðnÞþPFLCðnÞð8:26Þ

The ﬁrst component, PAVG, is responsible for providing the desired proﬁle (i.e., MG net power average) and is computed as follows:

1 XM PAVGðnÞ¼ PLGðn À kÞð8:27Þ M k¼1 where M is the number of samples in one day (i.e., window size of 24 h). The second component, PFLC, is used to modify the grid power profile according to the MG ERoC and the battery SOC. The ERoC, ṖAVG,isdefined as the first backward difference of the MG net power average, as follows (Arcos-Aviles et al. 2018): _ PAVGðnÞ¼½PAVGðnÞÀPAVGðn À 1Þ =Ts ð8:28Þ

The block diagram of the fuzzy ERoC strategy is shown in Fig. 8.3 where: Block 1 corresponds to a low-pass filter (LPF) used to obtain the MG net power average profile (8.27); Block 2 comprises a Derivative and Filter blocks (DF) used concurrently to obtain ṖAVG(n)(8.28) and limit the high-frequency gain and noise associated with the derivative term (Ang et al. 2005); Block 3 corresponds to the battery SOC estimator; and, Block 4 represents a FLC used for smoothing the power profile exchanged to the grid (Arcos-Aviles et al. 2018). Fuzzy logic controller design. All the design strategies of this chapter are based on a low complexity FLC of only two inputs and one output where the FLC assumes a Mamdani-based inference and a Center of Gravity defuzzification method (Passino and Yurkovich 1998). 178 D. Arcos-Aviles et al.

As it is well known, a FLC has different parameters, such as Membership Functions (MFs) number, type, mapping, and rule-base that, according to certain conditions, have to be adjusted to increase the controller effectiveness. Accordingly, the adjustment of fuzzy parameters of the energy management strategies described in this chapter is carried-out by an off-line adjustment procedure with the aim of minimizing the magnitude of the previously defined grid power quality criteria. This procedure uses the real generation and consumption data shown in Fig. 8.4 (i.e., one year of renewable generation and load demand power measured every 15 min in the MG installed at UPNa, Pamplona, Spain: 42°49′06″N 1°38′39″O) (Arcos-Aviles et al. 2017a, 2018) to train the FLC so that it can effectively respond to all possible scenarios of RES production and load consumption for this type of residential systems (Arcos-Aviles et al. 2017a). The adjustment procedure described in Arcos-Aviles et al. (2012) is outlined next and includes the following steps (Arcos-Aviles et al. 2018): Step 1: Set the initial FLC design. (1) Set the membership functions (MF) of inputs and outputs variables: number, type, and mapping. (2) Set the initial rule-base based on the heuristic knowledge about the MG behavior. Step 2: Adjust the MFs of inputs and outputs. By means of the real recorded data, the MFs parameters of inputs/output are adjusted to minimize the defined grid power profile quality criteria. Step 3: Optimize the initial rule-base. By using the real recorded data, the initial rule-base is adjusted to minimize the defined grid power profile quality criteria. As a result of the adjustment procedure, five triangular MFs are defined for each input variable, ṖAVG(n) and SOC(n), corresponding to five fuzzy subsets noted as NB, NS, ZE, PS, and PB where B stands for “Big”, S for “Small”, N for “Negative”, P for “Positive” and ZE for “Zero”. The MFs, shown in Fig. 8.5a, b, are distributed along the variation range of each input. Note that the variation range of the input SOC(n) has to satisfy the constraints defined by (8.13) and (8.14) whereas the variation range of the input ṖAVG(n)isdefined as follows: 8 A Review of Fuzzy-Based Residential Grid-Connected Microgrid … 179

(a) (b)

Fig. 8.4 Renewable power generation and load power demand measured at UPNa microgrid from July 2013 to July 2014 a photovoltaic power, b wind turbine power, c load power of the ﬁrst architecture, and d load power of the second architecture (Arcos-Aviles et al. 2017a)(©2017 Elsevier, Reprinted, with permission, from Arcos-Aviles et al. 2017a)

_ _ _ PAVG;MIN PAVGðnÞPAVG;MAX ð8:29Þ _ PAVG;MIN ¼ÀðÞÁ9=10 ðÞPWT =TW ð8:30Þ _ PAVG;MAX ¼ ðÞÁ9=10 ðÞPLOAD=TW ð8:31Þ where TW is the time window of one day and ṖAVG,MAX and ṖAVG,MAX are the maximum and minimum energy change in the MG, respectively, expressed in W/s according to the approximation developed in Marcos et al. (2014). Conversely, nine triangular MFs are deﬁned for the output variable, PFLC(n), which are associated to nine fuzzy subsets noted as NB, NM, NS, NSS, ZE, PSS, PS, PM and PB, where besides B, S, N, P and ZE previously deﬁned, M stands for “Medium” and SS for “Smallest”, as shown in Fig. 8.5c, where a negative MF implies a reduction in the grid power and a positive MF implies an increase in the 180 D. Arcos-Aviles et al.

(a) (b)

(c)

Fig. 8.5 Fuzzy ERoC strategy, adjusted MFs for a input ṖAVG(n), b input SOC(n), and c output PFLC(n) (Arcos-Aviles et al. 2018). ©2018 IEEE, Reprinted, with permission, from Arcos-Aviles et al. (2018) grid power. Similarly, the output MFs are distributed along the variation range deﬁned as follows (Arcos-Aviles et al. 2018):

Àk PFLCðnÞk ð8:32Þ being k the maximum power assigned by the fuzzy controller output (i.e., k = 1 kW). Finally, after the adjustment procedure the initial rule-base, which was built considering the linguistic knowledge about the MG behavior, is modiﬁed to obtain a set of rules which meet the strategy desired objectives. The ﬁnal rule-base consists in only 25 rules as presented in Table 8.1. For instance, the rule represented in the bold stands as: IF a high energy rate-of-change occurs in the MG due to an increase of PLG (i.e., ṖAVG(n) 0, ṖAVG is “PB”) AND if the energy stored in the ESS is very low (i.e., SOC(n) SOCMIN, SOC is “NB”) THEN the strategy strongly increases the power injection from the mains to charge the ESS (i.e., PFLC is “PB”). 8 A Review of Fuzzy-Based Residential Grid-Connected Microgrid … 181

Table 8.1 Adjusted PFLC(n) ṖAVG(n) rule-base for the FLC of the NB NS ZE PS PB fuzzy ERoC strategy (Arcos-Aviles et al. 2018) SOC(n) NB PSS PSS PS PM PB NS PSS ZE PSS PSS PM ZE NS NSS ZE PSS PS PS NSS NS NSS PM PSS PM NB NM NSS NSS NSS ©2018 IEEE, Reprinted, with permission, from Arcos-Aviles et al. (2018)

8.5.2 EMS Based on MG Generation and Demand Forecasting (EMS-FC)

The EMS-FC strategy was introduced in Arcos-Aviles et al. (2016b) and an extended version including experimental validation results was presented in Arcos-Aviles et al. (2017a). This strategy uses generation and demand forecasting to anticipate the future MG behavior. Then, according to the prediction error and the battery SOC the EMS-FC strategy performs the suitable control of the power exchanged to the gird. The EMS-FC strategy considers the ﬁrst power architecture shown in Fig. 8.1. Microgrid power forecast. Generation and demand forecasts are considered to obtain the forecast of the MG net power. Therefore, taking into account the mathematical models of PV and WT generators described in Sects. 8.3.1 and 8.3.2, respectively, the generation forecast is estimated by Numerical Weather Prediction (NWP) method (Yang et al. 2015; Foley et al. 2012) using the meteorological data provided by a local observatory (Meteogalicia 2018). The forecast of power generation, PGEN,FC, is expressed as follows (Arcos-Aviles et al. 2017a):

PGEN;FCðnÞ¼PPV;FCðnÞþPWT;FCðnÞð8:33Þ where PPV,FC is the PV power forecast (8.5) and PWT,FC is the WT power forecast (8.9). Moreover, due to the residence daily consumption pattern is similar from one day to the next one (excluding weekends and holidays) the load demand can be estimated by means of the persistence forecast model (Arcos-Aviles et al. 2016b). This model uses the past data as the forecast for the next time period (Hanna et al. 2014;Lü et al. 2014), so that, the load proﬁle for the next day will be the same as the previous one, as follows (Arcos-Aviles et al. 2017a):

PLOAD;FCðn þ M24Þ¼PLOADðnÞð8:34Þ 182 D. Arcos-Aviles et al. where PLOAD,FC is the forecast of load demand and M24 is the number of samples in the past 24-h. Note that in a real residential microgrid scenario, the daily consumption pattern between weekdays, weekends, and holidays is not the same, thus, the difference between them should be considered when computing the load demand forecast. Finally, the forecast of MG net power, PLG,FC(n), and the MG forecast error, PE(n), are computed as follows:

PLG;FCðnÞ¼PLOAD;FCðnÞÀPGEN;FCðnÞð8:35Þ

PEðnÞ¼PLGðnÞÀPLG;FCðnÞð8:36Þ

Control strategy. The block diagram of the EMS-FC is shown in Fig. 8.6 and includes a CMA filter block (Vamos and Craciun 2012) used to compute the MG net power average, a 3-h filter block used for both computing the MG average forecast error of the past 3-h and reducing the high variability associated with the forecast error, a FLC block used to smooth the grid power profile, and a battery control block which comprises a LPF used to compute the battery SOC average of the past 24-h and a battery SOC estimator. From Fig. 8.6, the grid power profile can be defined as the sum of three components (Arcos-Aviles et al. 2016b, 2017a):

PGRIDðnÞ¼PCTRðnÞþPSOCðnÞþPFLCðnÞð8:37Þ where PCTR(n) is the CMA filter block output, PSOC(n) is the battery control block output, and PFLC(n) is the FLC block output. These components are described next: The first component, PCTR(n), is computed by a CMA filter of 24-h, which estimates the MG net power average considering the MG net power average value

Fig. 8.6 Block diagram of the fuzzy EMS-FC strategy (Arcos-Aviles et al. 2016b, 2017a). ©2017 Elsevier, Reprinted, with permission, from Arcos-Aviles et al. (2017a), ©2016 IEEE, Reprinted, with permission, from Arcos-Aviles et al. (2016b) 8 A Review of Fuzzy-Based Residential Grid-Connected Microgrid … 183

12H of the past 12-h, PLG , and the MG net power forecast average value of the next 12H 12-h, PLG;FC, as follows (Arcos-Aviles et al. 2016b, 2017a): hi 1 P ðnÞ¼ Á P12HðnÞþP12H ðnÞ ð8:38Þ CTR 2 LG LG;FC

XM12 12Hð Þ¼ 1 ð À Þð: Þ PLG n PLG n k 8 39 M12 k¼1

XM12 12H ð Þ¼ 1 ð þ Þð: Þ PLG;FC n PLG;FC n k 8 40 M12 k¼1 being M12 the number of samples in 12-h. The second component, PSOC(n), is used by the strategy to keep the battery SOC center close to the 75% of the rated battery capacity (Pascual et al. 2015). This component is proportional to the error between the battery SOC reference value and the average battery SOC of the last 24-h, SOCREF and SOCAVG(n), respectively, as follows (Arcos-Aviles et al. 2016b, 2017a):

PSOCðnÞ¼ke Á ½ðSOCREF À SOCAVGðnÞ 8:41Þ

1 XM24 SOCAVGðnÞ¼ SOCðn À kÞð8:42Þ M24 k¼1 where ke is the proportional gain constant, which according to Pascual et al. (2015) is set to 0.05 kW/% to obtain a high enough phase margin in the battery control loop. Finally, the third component, PFLC, is used to improve the grid power proﬁle according to the battery SOC and the forecast error of the past 3-h. This component is computed by a FLC which considers the current battery SOC (8.10) and the MG average forecast error of the past 3-h deﬁned in (8.43), as follows (Arcos-Aviles et al. 2016b, 2017a):

XM3 3Hð Þ¼ 1 ð À Þð: Þ PE n PE n k 8 43 M3 k¼1 being M3 the number of samples of the past 3-h. Fuzzy logic controller design. Similarly to previous strategy based on MG ERoC, the FLC design follows the adjustment procedure described in Aviles et al. (2012). As a result, five triangular MFs are defined for each input variable, SOC(n) and 3H fi PE (n), and nine triangular MFs are de ned for the controller output, PFLC(n). These MFs (inputs and output) are distributed along the variation range defined in 184 D. Arcos-Aviles et al.

(a) (b) NB NS ZE PS PB NB NM NS NSS ZE PSS PS PM PB 1 1

0.8 0.8

0.6 0.6

0.4 0.4

0.2 0.2

0 0

3 111 1 3 P 0 P -P 1 0 1 +P N P P P P P P P e - P + P e 4 N 2 N 8 N 8 P 2 P 4 P 2 e 2 e

3H Fig. 8.7 Fuzzy EMS-FC strategy, adjusted MFs for a input PE (n) and b output PFLC(n) (Arcos-Aviles et al. 2016b). ©2017 Elsevier, Reprinted, with permission, from Arcos-Aviles et al. (2017a), ©2016 IEEE, Reprinted, with permission, from (Arcos-Aviles et al. 2016b)

(8.13), (8.44), and (8.45). Note that in this strategy Pe = 6 kW, PN = −0.3 kW and PP = 0.45 kW.

À 3Hð Þ ð : Þ Pe PE n Pe 8 44

PN PFLCðnÞPP ð8:45Þ

The MFs of the inputs are shown in Figs. 8.5b and 8.7a whereas the MFs of the output are presented in Fig. 8.7b. Finally, the adjusted rule-base (25 rules) is presented in Table 8.2. For instance, the rule represented in the bold cell stands as: IF the forecast error 3H in the MG is highly negative (PE (n) 0, the forecast of the MG net power is greater than the measured value) AND the energy stored in the ESS is very low (SOC(n) SOCREF which means that the battery is highly discharged) THEN the strategy strongly increases the grid power to charge the ESS (PFLC 0). Note that the rules over the diagonal imply that the power delivered by the mains increases

Table 8.2 Adjusted 3H PFLC(n) P (n) rule-base for the FLC of the E fuzzy EMS-FC strategy NB NS ZE PS PB Arcos-Aviles et al. (2017a) SOC(n)NBPB PB PSS PS PB NS PM NS PSS PS PS ZE PSS ZE ZE PSS PM PS NS PM NSS NS NSS PM NS NSS NSS NM NB ©2017 Elsevier, Reprinted, with permission, from Arcos-Aviles et al. (2017a) 8 A Review of Fuzzy-Based Residential Grid-Connected Microgrid … 185 according to the increase in the forecast error. Conversely, the rules below the diagonal indicate that the grid power injection decreases according to the increase of the energy stored in the battery (Arcos-Aviles et al. 2017a).

8.5.3 EMS Based on MG Energy Rate-of-Change Applied to an Electro-Thermal Microgrid (ERoC ETH-MG)

This strategy was presented in Arcos-Aviles et al. (2017b) and is carried out in the second power architecture under analysis (Fig. 8.2). In this architecture the main goal of the EMS strategy is to use the energy stored in the ESS to cover part of the load power required by the EWH to keep the water temperature in the hot water tank between established limits (i.e., from 45 to 65 °C). The MG net power for the second architecture is deﬁned as follows:

PLG ¼ PLOAD À PGEN þ PH;e ð8:46Þ

Control strategy. The design follows the control policies used in the strategy described in Sect. 8.5.1 and includes some additional statements to perform the suitable control of the EWH. The block diagram depicted in Fig. 8.3 is modiﬁed to include a EWH control block, as shown in Fig. 8.8, which is used to determine the amount of energy required by the electric heater that will be supplied by the ESS. This way, the total power required by the EWH will be handled by the ESS and the mains, reducing the MG net power (8.46) as follows (Arcos-Aviles et al. 2017b):

Ã ð Þ¼ ð ÞÀ B ð Þð: Þ PLG n PLG n PWH n 8 47 P ; ðnÞ; if SOCðnÞ [ SOC &P ; ðnÞ [ 0 PB ¼ H e R1 H e ð8:48Þ WH 0; otherwise

B where PWH is the power supplied by the battery for the EWH consumption and SOCR1 is a battery SOC threshold which allows power injection from the battery. fi Ã Note that since PLG is modi ed to PLG all variables depicted in Fig. 8.8 should be Ã fi fi re-computed considering the new PLG. For instance, the grid power pro le de ned in (8.26) now is modified as:

¼ Ã ð Þþ ð Þð: Þ PGRID PAVG n PFLC n 8 49

In contrast, the battery power is expressed as the sum of two components, as follows (Arcos-Aviles et al. 2017b): 186 D. Arcos-Aviles et al.

ð Þ¼ MG ð Þþ B ð Þð: Þ PBAT n PBAT n PWH n 8 50 MG ð Þ¼ ð ÞÀ ð Þð: Þ PBAT n PLG n PGRID n 8 51

Fuzzy logic controller design. After the adjustment procedure (Aviles et al. 2012), _ Ã the resulting MFs of input variables, PAVG(n) and SOC(n), are shown in Fig. 8.9a, Fig. 8.5b, respectively. The variation range of each input is defined by (8.52) and (8.13). Moreover, the MFs of the output, PFLC(n), are modified as shown in Fig. 8.9b whereas the output variation range is defined by (8.45) (i.e., _ PN = −1.4 kW and PP = 1.5 kW), being PA is the maximum variation of the derivative term defined in (8.31).

À _ Ã ð Þ ð : Þ PA PAVG n PA 8 52 8 A Review of Fuzzy-Based Residential Grid-Connected Microgrid … 187

(a) (b)

Table 8.3 Adjusted _ Ã PFLC(n) P (n) rule-base for the FLC of the AVG fuzzy ERoC ETH-MG NB NS ZE PS PB strategy Arcos-Aviles et al. SOC(n) NB PB PM PSS PM PB (2017b) NS PM PS PSS PS PM ZE ZE NSS ZE NSS NSS PS NM NSS NS NM NM PM NB NB NSS NM NB ©2017 IEEE, Reprinted, with permission, from Arcos-Aviles et al. (2017b)

The adjusted rule-base is presented in Table 8.3. Note that given the new objective of the EMS in the second architecture, the initial rule-base described in Table 8.1 is modiﬁed to the one shown in Table 8.3 to keep the ESS between secure limits and to smooth the grid power proﬁle under the new conditions presented in the electro-thermal MG.

8.5.4 EMS Based on MG Power Forecasting Applied to an Electro-Thermal Microgrid (EMS-FC ETH-MG)

The last EMS strategy deals with the analysis of the EMS-FC strategy when is applied to the second power architecture under study. As well as the previous strategy, the main goal of the EMS-FC ETH-MG strategy is to smooth the grid power proﬁle by means of performing the suitable control of the energy consumed by the EWH to keep the temperature in the hot water tank between 45 and 65 °C. 188 D. Arcos-Aviles et al.

Fig. 8.10 Block diagram of the fuzzy EMS-FC strategy applied to an electro-thermal MG

Control strategy. The block diagram of the EMS-FC ETH-MG strategy is shown in Fig. 8.10. As it can be seen, the block diagram comprises the same blocks described in Sect. 8.5.2 and includes the EWH control block previously described. Similarly, the use of the ESS to cover part of the power required by the EWH modiﬁes the MG net power as in (8.47) changing its forecast, and consequently, its corresponding prediction error, as follows:

Ã ð Þ¼ ð ÞÀ B ð Þð: Þ PLOAD n PLOAD n PWH n 8 53 Ã ð Þ¼ Ã ð ÞÀ ð Þð: Þ PLG;FC n PLOAD;FC n PGEN;FC n 8 54 Ã ð Þ¼ Ã ð ÞÀ Ã ð Þð: Þ PE n PLG n PLG;FC n 8 55

Note that the renewable power generation is not affected by the use of the EWH, thus, the forecast of power generation, PGEN,FC, is maintained (Arcos-Aviles 2016a). Fuzzy logic controller design. The MFs for input variables are the same to those 3HÃ presented in Fig. 8.7a (input PE ), Fig. 8.5b (input SOC), and Fig. 8.9b (output 8 A Review of Fuzzy-Based Residential Grid-Connected Microgrid … 189

Table 8.4 Adjusted 3HÃ PFLC(n) P (n) rule-base for the FLC of the E fuzzy EMS-FC ETH-MG NB NS ZE PS PB strategy (Arcos-Aviles 2016a) SOC(n)NBPBPMPSPMPB NS PM PS PSS PS PM ZE NS ZE ZE PSS NSS PS NM NS NSS NS NM PM NB NSS NM NM NB

PFLC). Note that the adjustment procedure establishes the variation range of the output variable between PN = −0.8 kW and PP = 1.35 kW. Finally, the resulting rule-base is presented in Table 8.4.

8.6 Results and Comparison

The simulation of the EMSs described in this chapter are accomplished using the historical data recorded from July 2013 to July 2014 in the MG installed at UPNa (Pamplona, Spain: 42°49′06″N 1°38′39″O). The results and comparisons are performed by means of numerical simulations using Matlab®. A baseline of comparison is obtained considering the MG net power profile, shown in Fig. 8.11, and its quality criteria for each architecture under analysis (i.e., considering a MG scenario without ESS and EMS, PBAT = 0; therefore, according to (8.1) PGRID = PLG for the first architecture and according to (8.4) PGRID = PLOAD − PGEN + PH,e for the second architecture) as presented in Table 8.5. Figures 8.12 and 8.13 show the grid power profile and the battery SOC evolution achieved by the fuzzy ERoC strategy and fuzzy EMS-FC strategies, when they are

(a) (b)

Fig. 8.11 Baseline grid power proﬁles of the ﬁrst and second architectures (a and b), respectively 190 D. Arcos-Aviles et al.

Table 8.5 Grid power proﬁle quality criteria baseline values

EMS strategy PG,MAX (kW) PG,MIN (kW) MPD APD PPV (W/h) (W/h)

PLG electric MG 5.75 −6.45 18,468 1121 13.3

PLG electro-thermal MG 6.53 −6.45 18,468 1221 5.99

(a) (b)

Fig. 8.12 Comparison between the fuzzy ERoC and the fuzzy ERoC ETH-MG strategies: Grid power profile (a, b), Battery SOC (c, d) applied to both architectures under analysis (i.e., electric and electro-thermal MGs). As it can be seen, the four analyzed strategies improve the grid power profile with respect to the baseline profiles shown in Fig. 8.11. The strong fluctuations and power peaks in the grid power are minimized. Moreover, the correct battery SOC evolution is confirmed when observing that the battery SOC is kept between secure limits in all the strategies described in this chapter. Note that since the architectures under study considers an ESS based on lead-acid battery bank, a proper battery SOC evolution occurs when the battery SOC fluctuates close to the 75% of the rated battery capacity (Figs. 8.12c, d and 8.13c, d). 8 A Review of Fuzzy-Based Residential Grid-Connected Microgrid … 191

(a) (b)

Fig. 8.13 Comparison between the fuzzy EMS-FC and the fuzzy EMS-FC ETH-MG strategies: Grid power proﬁle (a, b), Battery SOC (c, d)

Although the results achieved through the fuzzy ERoC strategy, for both power architectures, are fairly acceptable, the fuzzy EMS-FC strategy, in both architectures, reduces the power fluctuations and power peaks in the grid power profile, as shown in Fig. 8.14. This improvement is due to the use of generation and demand forecasts to anticipate the energy changes that occur in the MG. The controller ability to predict future events is evidenced in Fig. 8.15, which compare for the electric microgrid the features of the fuzzy ERoC and the fuzzy EMS-FC strategies, along four consecutive days of the year under study. As it can be seen, at points A and C in Fig. 8.15, the EMS-FC strategy predicts that there will be an increase in generation on the next day (17/04/2014 and 19/04/ 2017, respectively) close to 14:00 P.M. (i.e., PLG profile will be negative due to the increase in power generation, black dotted line). Consequently at these points, the EMS-FC strategy decides to decrease the grid power (blue solid line) yielding the battery discharging (green solid line) and preparing the ESS so that it can absorb the future power generation. Conversely, the grid power profile achieved by the ERoC strategy (red dashed line), where power forecast is not considered, is kept constant 192 D. Arcos-Aviles et al.

(a) (b)

Fig. 8.14 Grid power proﬁle comparison between the fuzzy ERoC and the fuzzy EMS-FC strategies

Fig. 8.15 Grid power profile and battery SOC comparison between the fuzzy ERoC and the fuzzy EMS-FC strategies (Arcos-Aviles et al. 2017a). ©2017 Elsevier, Reprinted, with permission, from Arcos-Aviles et al. (2017a) at these points (A and C) and rapidly falls as the power generation increases (i.e., negative PLG profile) producing the battery charging (magenta dash-dot line) over 90% of its rated capacity and a negative power peak in the grid power profile, as it can be seen at points B and D in Fig. 8.15. The experimental validation of the fuzzy ERoC and the fuzzy EMS-FC strategies can be found in Arcos-Aviles et al. (2018) and (2017a), respectively. 8 A Review of Fuzzy-Based Residential Grid-Connected Microgrid … 193

(a) (b)

Fig. 8.16 Distribution of the energy required by the EWH a ERoC ETH-MG strategy and b EMS-FC ETH-MG strategy

Table 8.6 Grid power proﬁle quality criteria comparison for the Fuzzy-based energy management strategies addressed in this chapter. Bold contents correspond to baseline values

EMS strategy PG,MAX PG,MIN MPD APD PPV (kW) (kW) (W/h) (W/h)

PLG electric MG 5.75 −6.45 18,468 1121 13.3 Fuzzy ERoC (Arcos-Aviles et al. 1.83 −2.04 817 56.15 2.79 2018) Fuzzy EMS-FC (Arcos-Aviles et al. 1.89 −1.48 480 51.79 2.76 2017a)

PLG electro-thermal MG 6.53 −6.45 18,468 1221 5.99 Fuzzy ERoC ETH-MG 2.95 −2.31 1052 75.67 1.26 (Arcos-Aviles et al. 2017b) Fuzzy EMS-FC ETH-MG 2.56 −1.89 846 75.13 1.26 (Arcos-Aviles et al. 2016a)

Regarding to the electro-thermal MG architecture, the suitable use of the energy stored in the battery to supply part of the energy required by the EWH leads to minimize the amount of energy coming form the mains required for the DHW consumption. Figure 8.16 shows the distribution of the energy used by the EWH to keep the water temperature in the hot water tank between 45 and 65 °C. The ESS significantly contributes with the energy required by the EWH. Note that the energy provided by the ESS represents the energy saved by the MG. Finally, the proper behavior of all strategies described in this chapter is also verified through the defined grid power profile quality criteria, which are summarized in Table 8.6. In short, all the analyzed fuzzy strategies achieve an important reduction in the quality criteria with respect the baseline values. Table 8.6 highlights that for each of the microgrid architectures, the use of forecast data and the additional degree of freedom provided by the electro-thermal microgrid improve all the grid power profile quality criteria (with respect to the baseline values of each 194 D. Arcos-Aviles et al. architecture). The choice of one of the strategies presented in this chapter evidently depend on the architecture (i.e., what are the controllable elements) and if forecasted data are available or not.

8.7 Conclusion

This chapter has addressed the design of different energy management strategies based on Fuzzy Logic Control for a grid connected residential microgrid with solar and wind renewable sources. The analysis has considered two different power architectures namely, only electrical and electro-thermal residential microgrids. The first architecture only includes the renewable generation and domestic load demand of the residence. Conversely, the electro-thermal microgrid has considered, beside all the elements of the electrical microgrid, the thermal requirements, i.e., electric water heater, solar thermal collectors, water storage tank, and domestic hot water consumption of the residence. The designs presented in this chapter have been developed using real recorded data of electrical power generation and consumption from July 2013 to July 2014 of the residential microgrid under study. The main objective selected for the design of the energy management strategy of these architectures has been the reduction of the power fluctuations in the power exchanged with the grid, while keeping the energy storage system within secure limits to preserve its life. Reaching this objective will facilitate the integration of the RES into the mains and contribute to the reduction of both the user electricity bill and the mains overload. The preliminary step of the designs has been to define a set of quality criteria evaluating the grid power profile behavior. By this way, the features of an EMS design can be quantified in terms of power fluctuations since an improved design results in lower values of each quality criterion. These quality criteria were additionally used to adjust the parameters of the FLC presented in this work by an off-line learning process using the real recorded data referred above. The first fuzzy EMS design is carried out by means of a two-input, one output, and 25-rules low complexity FLC, which quantify the magnitude of the energy changes in the MG, and in accordance with these changes the EMS increases, decreases, or maintains the power delivered/absorbed by the mains. A set of numerical simulations has shown that this design allows the EMS to react quickly against the MG energy changes to set the battery SOC close to the 75% of the rated battery capacity, so that the battery available dynamic range can compensate the fluctuations in the MG net power, smoothing the grid power profile and keeping the battery SOC between secure limits. This chapter has also faced the FLC design when power forecast data are available. A new FLC design referred as EMS-FC includes the MG power forecast to predict its future behavior, so that, the forecast error and the battery SOC are used by the FLC to modify the grid power profile. The fuzzy EMS-FC strategy computes the grid power as the sum of three components: the first one is used to provide the 8 A Review of Fuzzy-Based Residential Grid-Connected Microgrid … 195 desired MG average profile through a CMA filter of 24-h; the second one is used to keep the battery SOC close to the 75% of the rated battery capacity; and, the third one is used to smooth the grid power profile according to the forecast error and the battery SOC. The simulation results and further comparison proved the feasibility of this strategy which concurrently holds the SOC of the battery close to the 75% of the rated battery capacity and minimizes the grid power fluctuations with respect to previous strategies. The enhanced fuzzy-based EMS strategies were also applied to electro-thermal MG architecture. The results have demonstrated that the increase in freedom degrees in the MG, i.e., adding controllable elements in the MG (thermal elements) can facilitate the grid power profile control. The use of the EWH and the hot water tank have allowed the improvement of the power exchanged with the grid and the reduction of the amount of power supplied by the mains required to meet the domestic hot water consumption, which implies a cost reduction for the user. It can be conclude that this chapter has evidenced the feasibility of using low complexity FLC of only 25 rules for the EMS strategy design in electric and electro-thermal MG architectures to smooth the grid power profile while preserving the energy storage system operation within secure limits. FLC allowed transforming the heuristic knowledge of the MG desired behavior of previous works in the form of linguistic rules, making easier their understanding and facilitating the EMS design. Even not addressed in this chapter, experimental validations can be found in Arcos-Aviles et al. (2017a, 2018). It can be finally pointed out that the energy management for a grid-connected microgrid is still a topic of ongoing research. The following aspects for improving both the reliability of the microgrid and the energy management strategy design can be addressed in further studies, namely: • The analysis of the EMS strategy under a microgrid scenario including more freedom degrees, such as, controllable loads, would allow the integration of a Demand Side Management (DSM) or Demand Respond (DR) techniques, which would improve the system performance. • The addition of multiple storage elements in the microgrid scheme would allow the integration of a multi-agent based management to perform a MG cooperative control. • The expansion of the EMS strategy design for the case of multiple intercon- nected microgrids would allow sharing the energy requirements of each microgrid, so that, they can be supplied for another microgrid, with the aim of improving the MGs performance at the local and global level. • Additional technical aspects could be considered for the estimation of the state-of-charge of the energy storage system, for instance, number of charge-discharge cycles, and charging/discharging power rates and limits, which would help to preserve the storage system lifetime (i.e., battery ageing issues). 196 D. Arcos-Aviles et al.

• The parameter optimization of the fuzzy logic controller developed in this chapter could be compared with the optimization accomplished through sophisticated algorithms, for instance, Particle Swarm Optimization (PSO) and Cuckoo optimization algorithms. • The feasibility of this study could be further improved by means of considering a new cost function, to include all the beneﬁts of the deﬁned quality criteria and additionally as well as electricity tariffs with the purpose of minimizing the operational cost of the microgrid. • The use of other techniques such as Model Predictive Control (MPC) should be analyzed and compared with the approach presented in this chapter.

References

Ally C, Bahadoorsingh S, Singh A, Sharma C (2015) A review and technical assessment integrating wind energy into an island power system. Renew Sustain Energy Rev 51:863–874 Ang KH, Chong G, Li Y (2005) PID control system analysis, design, and technology. IEEE Trans Control Syst Technol 13(4):559–576 Anuphappharadorn S, Sukchai S, Sirisamphanwong C, Ketjoy N (2014) Comparison the economic analysis of the battery between lithium-ion and lead-acid in PV stand-alone application. Energy Procedia 56:352–358 Arcos-Aviles D (2016a) Energy management strategies based on fuzzy logic control for grid-tied domestic electro-thermal microgrid. Universitat Politècnica de Catalunya Arcos-Aviles D, Guinjoan F, Barricarte J, Marroyo L, Sanchis P, Valderrama H (2012) Battery management fuzzy control for a grid-tied microgrid with renewable generation. In: IECON 2012—38th annual conference on IEEE industrial electronics society, pp 5607–5612 Arcos-Aviles D, Vega C, Guinjoan F, Marroyo L, Sanchis P (2014a) Fuzzy logic controller design for battery energy management in a grid connected electro-thermal microgrid. In: 2014 IEEE 23rd international symposium on industrial electronics (ISIE), pp 2014–2019 Arcos-Aviles D, Espinosa N, Guinjoan F, Marroyo L, Sanchis P (2014b) Improved fuzzy controller design for battery energy management in a grid connected microgrid. In: IECON 2014—40th annual conference of the IEEE industrial electronics society, pp 2128–2133 Arcos-Aviles D, Pascual J, Marroyo L, Sanchis P, Guinjoan F, Marietta MP (2015) Optimal fuzzy logic EMS design for residential grid-connected microgrid with hybrid renewable generation and storage. In: 2015 IEEE 24th international symposium on industrial electronics (ISIE), pp 742–747 Arcos-Aviles D, Guinjoan F, Marietta MP, Pascual J, Marroyo L, Sanchis P (2016b) Energy management strategy for a grid-tied residential microgrid based on Fuzzy Logic and power forecasting. In: IECON 2016—42nd annual conference of the IEEE industrial electronics society, pp 4103–4108 Arcos-Aviles D, Pascual J, Guinjoan F, Marroyo L, Sanchis P, Marietta MP (2017a) Low complexity energy management strategy for grid proﬁle smoothing of a residential grid-connected microgrid using generation and demand forecasting. Appl Energy 205:69–84 Arcos-Aviles D, Sotomayor D, Proano JL, Guinjoan F, Marietta MP, Pascual J, Marroyo L, Sanchis P (2017b) Fuzzy energy management strategy based on microgrid energy rate-of-change applied to an electro-thermal residential microgrid. In: 2017 IEEE 26th international symposium on industrial electronics (ISIE), pp 99–105 8 A Review of Fuzzy-Based Residential Grid-Connected Microgrid … 197

Arcos-Aviles D, Pascual J, Marroyo L, Sanchis P, Guinjoan F (2018) Fuzzy logic-based energy management system design for residential grid-connected microgrids. IEEE Trans Smart Grid 9 (2):530–543 Asmus P, Lauderbaugh A, Lawrence M (2013) Market data: microgrids. Campus / Institutional, Military, and Remote Microgrids Barricarte JJ, Martín IS, Sanchis P, Marroyo L (2011) Energy management strategies for grid integration of microgrids based on renewable energy sources. In: 10th International conference on sustainable energy technologies, pp 4–7 Black J, Larson R (2007) Strategies to overcome network congestion in infrastructure systems. J Ind Syst Eng 1(2):97–115 Chen Y-H, Lu S-Y, Chang Y-R, Lee T-T, Hu M-C (2013) Economic analysis and optimal energy management models for microgrid systems: a case study in Taiwan. Appl Energy 103:145–154 Chong WT, Hew WP, Yip SY, Fazlizan A, Poh SC, Tan CJ, Ong HC (2014) The experimental study on the wind turbine’s guide-vanes and diffuser of an exhaust air energy recovery system integrated with the cooling tower. Energy Convers Manag 87:145–155 Comodi G, Giantomassi A, Severini M, Squartini S, Ferracuti F, Fonti A, Nardi Cesarini D, Morodo M, Polonara F (2015) Multi-apartment residential microgrid with electrical and thermal storage devices: experimental analysis and simulation of energy management strategies. Appl Energy 137:854–866 Danish Wind Industry Association (2003) Wind energy reference manual part 1: wind energy concepts. http://drømstørre.dk/wp-content/wind/miller/windpower web/en/stat/unitsw. htm#roughness European Commission (2018) Clean energy for all Europeans. https://ec.europa.eu/energy/en/ topics/energy-strategy-and-energy-union/clean-energy-all-europeans. Accessed 11 Aug 2018 Fathima AH, Palanisamy K (2015) Optimization in microgrids with hybrid energy systems—a review. Renew Sustain Energy Rev 45:431–446 Foley AM, Leahy PG, Marvuglia A, McKeogh EJ (2012) Current methods and advances in forecasting of wind power generation. Renew Energy 37(1):1–8 Fossati JP, Galarza A, Martín-Villate A, Echeverría JM, Fontán L (2015) Optimal scheduling of a microgrid with a fuzzy logic controlled storage system. Int J Electr Power Energy Syst 68:61–70 Guasch D, Silvestre S (2003) Dynamic battery model for photovoltaic applications. Prog Photovolt Res Appl 11(3):193–206 Hanna R, Kleissl J, Nottrott A, Ferry M (2014) Energy dispatch schedule optimization for demand charge reduction using a photovoltaic-battery storage system with solar forecasting. Sol Energy 103:269–287 Hatziargyriou N (2014) Microgrids: architectures and control. Wiley, Chichester, UK Kim Seul-Ki, Jeon Jin-Hong, Cho Chang-Hee, Ahn Jong-Bo, Kwon Sae-Hyuk (2008) Dynamic modeling and control of a grid-connected hybrid generation system with versatile power transfer. IEEE Trans Ind Electron 55(4):1677–1688 Lasseter RH (2002) MicroGrids. IEEE Power Eng Soc Winter Meet 1:305–308 Lorenzo E (2011) Energy collected and delivered by PV modules. In: Luque A, Hegedus S (eds) Handbook of photovoltaic science and engineering. Wiley, Chichester, UK, pp 984–1042 Lü X, Lu T, Kibert CJ, Viljanen M (2014) A novel dynamic modeling approach for predicting building energy performance. Appl Energy 114:91–103 Mahmud MA, Hossain MJ, Pota HR, Nasiruzzaman ABM (2011) Voltage control of distribution networks with distributed generation using reactive power compensation. In: IECON 2011— 37th annual conference of the IEEE industrial electronics society, pp 985–990 Manwell JF, McGowan JG, Rogers AL (2009) Wind energy explained: theory, design and application. Wiley, Chichester, UK, pp 23–87 Marcos J, de la Parra I, García M, Marroyo L (2014) Control strategies to smooth short-term power ﬂuctuations in large photovoltaic plants using battery storage systems. Energies 7(10):6593– 6619 198 D. Arcos-Aviles et al.

Masters CL (2002) Voltage rise: the big issue when connecting embedded generation to long 11 kV overhead lines. Power Eng J 16(1):5–12 Mathew S (2006) Wind energy: fundamentals, resource analysis and economics. Springer, Berlin, pp 11–88 Meteogalicia. Servidor THREDDS de MeteoGalicia. http://www.meteogalicia.es/web/index.action. Accessed 05 July 2018 Mohamed A, Mohammed O (2013) Real-time energy management scheme for hybrid renewable energy systems in smart grid applications. Electr Power Syst Res 96:133–143 Niknam T, Azizipanah-Abarghooee R, Narimani MR (2012) An efficient scenario-based stochastic programming framework for multi-objective optimal micro-grid operation. Appl Energy 99:455–470 Olatomiwa L, Mekhilef S, Ismail MS, Moghavvemi M (2016) Energy management strategies in hybrid renewable energy systems: a review. Renew Sustain Energy Rev 62:821–835 Parisio A, Rikos E, Tzamalis G, Glielmo L (2014) Use of model predictive control for experimental microgrid optimization. Appl Energy 115:37–46 Parissis O-S, Zoulias E, Stamatakis E, Sioulas K, Alves L, Martins R, Tsikalakis A, Hatziargyriou N, Caralis G, Zervos A (2011) Integration of wind and hydrogen technologies in the power system of Corvo island, Azores: a cost-benefit analysis. Int J Hydrogen Energy 36 (13):8143–8151 Pascual J, Sanchis P, Marroyo L (2014) Implementation and control of a residential electrothermal microgrid based on renewable energies, a hybrid storage system and demand side management. Energies 7(1):210–237 Pascual J, Barricarte J, Sanchis P, Marroyo L (2015) Energy management strategy for a renewable-based residential microgrid with generation and demand forecasting. Appl Energy 158:12–25 Passino K, Yurkovich S (1998) Fuzzy control. Addisson-Wesley, Menlo Park, CA Schnitzer D, Lounsbury DS, Carvallo JP, Deshmukh R, Apt J, Kammen DM (2014) Microgrids for rural electrification: a critical review of best practices based on seven case studies Serraji M, Boumhidi J, Nfaoui EH (2015) MAS energy management of a microgrid based on fuzzy logic control. Intell. Syst. Comput. Vis. (ISCV) 2015:1–7 Shinji T, Sekine T, Akisawa A, Kashiwagi T, Fujita G, Matsubara M (2008) Reduction of power fluctuation by distributed generation in micro grid. Electr Eng Japan 163(2):22–29 Tascikaraoglu A, Boynuegri AR, Uzunoglu M (2014) A demand side management strategy based on forecasting of residential renewable sources: a smart home system in Turkey. Energy Build 80:309–320 Tazvinga H, Zhu B, Xia X (2015) Optimal power flow management for distributed energy resources with batteries. Energy Convers Manag 102:104–110 Tuballa ML, Abundo ML (2016) A review of the development of Smart Grid technologies. Renew Sustain Energy Rev 59:710–725 Vamos C, Craciun M (2012) Noise Smoothing. In: Vamos C, Craciun M (eds) Automatic trend estimation. Springer Netherlands, Dordrecht, pp 43–59 Velik R, Nicolay P (2014) Grid-price-dependent energy management in microgrids using a modified simulated annealing triple-optimizer. Appl Energy 130:384–395 Xue X, Wang S, Sun Y, Xiao F (2014) An interactive building power demand management strategy for facilitating smart grid optimization. Appl Energy 116:297–310 Yang C, Thatte AA, Xie L (2015) Multitime-scale data-driven spatio-temporal forecast of photovoltaic generation. IEEE Trans Sustain Energy 6(1):104–112 Yoo J, Park B, An K, Al-Ammar EA, Khan Y, Hur K, Kim JH (2012) Look-ahead energy management of a grid-connected residential PV system with energy storage under time-based rate programs. Energies 5(12):1116–1134 8 A Review of Fuzzy-Based Residential Grid-Connected Microgrid … 199

Zhao B, Zhang X, Chen J, Wang C, Guo L (2013) Operation optimization of standalone microgrids considering lifetime characteristics of battery energy storage system. IEEE Trans Sustain Energy 4(4):934–943 Zhou H, Bhattacharya T, Tran D, Siew TST, Khambadkone AM (2011) Composite energy storage system involving battery and ultracapacitor with dynamic energy management in microgrid applications. IEEE Trans Power Electron 26(3):923–930 Chapter 9 Analyzing Alternative Energy Mutual Fund Performance in the Spanish Market

Carmen-Pilar Martí-Ballester

Abstract Investors are becoming aware of the important role of renewable energy sources for mitigating global warming, which has encouraged them to invest in alternative energy mutual funds. However, the effectiveness of such financial instruments to mobilize private investments depends on the ability of managers to increase the investors’ wealth. For this reason, this paper examines the financial performance of alternative energy mutual funds compared to conventional and thematic market benchmarks in the Spanish market. To do so, we use a sample of 42 alternative energy mutual funds. Using these sample data, we implement a Carhart (1997) four-factor model. The results show that the use of a renewable energy index or conventional indexes affects fund performance. 19.05% of alternative energy mutual funds significantly exceed the renewable energy index, while 80.95% of alternative energy funds perform similarly to the market. There is no evidence of any effect of size or operating costs on the financial performance of funds.

Keywords Alternative energy mutual funds Á Conventional benchmark Alternative energy benchmark Á Size effect Á Total expense effect

9.1 Introduction

The use of energy from fossil fuels produces greenhouse gas emissions that contribute to global warming and climate change. Stakeholders concerned about environmental issues are putting pressure on ﬁrms to change to sustainable energy practices. The European Union is leading the way by promoting the Policy Framework for Climate and Energy, which encourages member states to meet its target of a 20% share of renewable energy sources in gross ﬁnal energy consumption—enacting the renewable energy directive (2009/28/EC)—and of 20%

C.-P. Martí-Ballester (&) Universitat Autònoma de Barcelona, 08193 Bellaterra, Spain e-mail: [email protected]

© Springer Nature Singapore Pte Ltd. 2019 201 E. Motoasca et al. (eds.), Energy Sustainability in Built and Urban Environments, Energy, Environment, and Sustainability, https://doi.org/10.1007/978-981-13-3284-5_9 202 C.-P. Martí-Ballester energy efficiency improvement by 2020 based on 2007 projections—enacting the energy efficiency (2012/27/EU) directive, respectively, by providing longer term visibility and investment security to firms that adopt RES projects. If it reaches both of these values, the European Union will have reduced its greenhouse gas emissions by 20% by 2020 compared with 1990 levels (Organization for Economic Co-operation and Development 2014). The governments of EU countries have endorsed these proposals and established their own national renewables and energy efficiency targets. Specifically, the Spanish government has committed to reaching similar objectives to the EU global target, increasing its share of renewable energies to 20% of gross final energy consumption by 2020. The actions taken to meet these renewable targets are specified in the national renewable energy action plan (RENAP) for the 2011–2020 period in accordance with the 2009/28/EC Directive. These actions are focused on (1) promoting the development of renewable technologies related to wind, solar thermal electricity, photovoltaic and biomass energy, (2) monitoring and controlling renewable energy sources through the Control Centre of Renewable Energies (CECRE), (3) supporting schemes for renewables, and (4) promoting green consumer engagement. As a consequence of these policies, the evolution of renewable energy contributions and CO2 emissions has improved in relation to 2004, as shown in Fig. 9.1.

Fig. 9.1 Evolution of renewable energy contribution and CO2 emissions. Source Eurostat. CO2 emissions in millions of tons 9 Analyzing Alternative Energy Mutual Fund … 203

Research and development efforts on renewable energy have allowed Spanish firms to assume leading positions around the world in wind, solar thermal, and photovoltaic energy production. These firms benefit from public support schemes that make renewable energy technologies competitive and avoid distorting energy prices and the energy market (Ruiz-Romero et al. 2012). However, the financial support from public entities remains insufficient to achieve the decarbonization objectives in the long term, and new financial measures and mechanisms are needed to ensure the active participation of private sector financial institutions as pointed out in the 22nd Conference of the Parties (COP22) to the United Nations Framework Convention on Climate Change in Marrakech in 2016. Private and public financial support should be combined to reduce project risks. The development of alternative energy technologies in the early stages has been supported by European governments to mitigate high initial cost and financing risks and to compensate low benefits of firms. This public funding should phase out as renewable energy projects mature and its lower risk can be assumed by the private sector returning proportional benefits to investors. The level of maturity favors the mobilization of private finance coming from financial institutions to meet the long-term Spanish targets for renewable energy (Organization for Economic Co-operation and Development 2014). Alternative energy mutual funds, as a financial institution, could be a powerful financial instrument to channel private finance into renewable energy-related projects because they enable (1) investors with long-term investment horizons to invest in a globally diversified and profes- sionally managed portfolio, which considers their renewable energy principles without sacrificing their long-term financial performance, and therefore, they contribute to energy sustainability and (2) to firms to meet their long-term financing needs to implement renewable energy practices contributing to energy sustainability (Paetzold and Busch 2014). Given that Spanish investors are becoming aware of the need to use renewable energies as a consequence of the actions of the national government to promote green consumer engagement, they could be willing to invest in alternative energy mutual funds if fund managers are able to provide them with high long-term risk-adjusted returns. For this reason, this paper analyzes the financial performance of alternative energy mutual funds. From modern portfolio theory (Markowitz 1952), the adoption of alternative energy criteria in the portfolio selection process could negatively influence the risk-adjusted returns of mutual funds and therefore the investors’ wealth due to alternative energy mutual funds only investing in firms that implement renewable energy practices, which reduces their investment opportunities while increasing the diversifiable risk of the portfolio. According to this theory, alternative energy mutual funds underperform the conventional market benchmark (a fully diversified portfolio). Additionally, the implementation of renewable energy practices requires high initial investments with a long-term payback, which negatively affects the short-term corporate financial performance from neoclassical economic theory 204 C.-P. Martí-Ballester

(Friedman 1970). However, the integration of renewable technologies in the production process allows firms to reduce greenhouse gas emission by avoiding pollution risks and to offer green certificated products to their customers. This leads firms to reduce their costs while improving their reputation, which positively affects long-term corporate financial performance from stakeholder theory (Freeman 1984). Alternative energy mutual funds could benefit from investing in firms that implement renewable energy activities in the long term, so we hypothesize that alternative energy mutual funds outperform (conventional and specialized) the market benchmark in the Spanish market. Previous literature examining alternative energy fund performance is scarce. The only study to explicitly analyze this issue is Reboredo et al. (2017). They particularly examine the financial performance of alternative energy mutual funds in relation to corporate mutual funds implementing the multifactor model proposed by Bollen and Busse (2001) for the 42 renewable mutual funds for the 2010–2016 period. Their findings indicate that green investors are paying a premium for investing in renewable energy funds. In this study, Reboredo et al. (2017) adopt the MSCI AC Europe TR index for the European market and the Fama and French index for the US market as conventional market benchmarks. However, the investment style of the conventional market benchmark could be different to that of alternative energy mutual funds (Martí-Ballester 2017; Bauer et al. 2005). Extending on the study by Reboredo et al. (2017) we analyze the effect of thematic and conventional market benchmarks on the risk-adjusted returns of alternative energy mutual funds at the individual fund level. The analysis of individual fund performance could be important for investors because it allows them to identify whether any individual fund outperforms its peers or is able to surpass the market benchmark. Additionally, Reboredo et al. (2017) analyze the effect of mutual fund characteristics on financial fund performance at the aggregate fund level. Renewable energy mutual funds that accumulate large amounts of assets could benefit from the existence of economies of scales that enable managers to reduce costs by improving the financial performance of funds (Ferreira et al. 2013). High management costs could erode the fund’s financial performance due to the existence of inefficient procedures (Ferreira et al. 2013; Malkiel 1995; Carhart 1997). Extending on the study by Reboredo et al. (2017), we analyze the effect of mutual fund characteristics on the risk-adjusted returns of alternative energy mutual funds at the individual fund level. The rest of this paper is structured as follows. Section 9.2 sets out the methodology used to evaluate the financial performance of alternative energy mutual funds. Section 9.3 provides information on the data used. Section 9.4 presents the empirical results. Section 9.5 summarizes the main results and presents some concluding remarks. 9 Analyzing Alternative Energy Mutual Fund … 205

9.2 Methodology

To examine the ﬁnancial performance of alternative energy mutual funds, we estimate the universally accepted Carhart (1997) four-factor model. This model reveals the fund’s excess return in relation to the market return for a given level of risk while it controls the impact of investment styles on risk-adjusted return by incorporating the size, book-to-market ratio and momentum strategy factors in the traditional Jensen (1968) model. Alternative energy funds tend to invest in (1) medium-large stocks, according to the Morningstar database, which could be more illiquid than small stocks that negatively affect fund performance and (2) growth stocks which have higher return volatility than value stocks leading to inferior risk-adjusted return (Ibikunle and Steffen 2017). This high return volatility of the growth stocks could affect the momentum strategy as indicated by Ibikunle and Steffen (2017). We present this model in the following expression: ÀÁ À ¼ a þ b À þ b þ b þ b þ l ð : Þ ra;t ri;t 0a 0a rm;t ri;t 1aSMBt 2aHMLt 3aMOMt a;t 9 1 where ra,t is the return on an equally weighted portfolio of funds in day t obtained from Thomson Reuters EIKON/Datastream database. ri,t is the daily return of the Euro Repo Index as the risk-free rate at time t, obtained from NEX Data. a0a represents the four-factor-adjusted return of portfolio a, a positive (negative) and signiﬁcant value indicating that the fund outperforms (underperforms) the market benchmark. rm,t represents the return on the conventional stock market index (MSCI World 1200 Index and S&P Global 1200 Index) and the sector stock market index (S&P Global Clean Energy Index) using as a proxy the market return at time t, obtained from Thomson Reuters EIKON/Datastream database. SMBt represents small-scale risk exposure using as a proxy the differential return between a small capitalization portfolio (MSCI World Small Cap Index) and a large capitalization portfolio (MSCI World Large Cap Index) at time t, obtained from Thomson Reuters EIKON/Datastream database. HMLt represents the bankruptcy risk using as a proxy the differential return between a value portfolio (MSCI World Value Index) and a growth portfolio (MSCI World Growth Index) at time t, obtained from Thomson Reuters EIKON/Datastream database. MOMt represents the momentum risk factor using as a proxy the return of the MSCI World Momentum Index at time t which is calculated taking into account the return difference between winner portfolio and loser portfolio. This index is obtained from Thomson Reuters EIKON/Datastream database. 206 C.-P. Martí-Ballester

b0a, b1a, b2a, b3a represent the factor loadings of the four style factors. la,t is the idiosyncratic (diversifiable) portfolio risk at time t. This risk is mitigated by diversifying one’s portfolio. We calculate the return for mutual funds as the variation relative to the daily net asset value expressed in euros, while the return for S&P and MSCI indexes (S&P Global Clean Energy Index, MSCI World 1200 Index, S&P Global 1200 Index, MSCI World Small Cap Index, MSCI World Large Cap Index, MSCI World Value Index, MSCI World Growth Index, and MSCI World Momentum Index) is given as the variation relative to the daily price index expressed in euros. The standard errors for the regression coefficients are estimated by implementing the Newey–West (1987) procedure. To compare the risk-adjusted returns obtained by using specialized and conventional global market indexes as benchmarks, we use the t-Student test. As a dependent variable we use the risk-adjusted return while our independent variable is (1) a dummy variable indicating if the market benchmark is a specialized index (S&P Global Clean Energy Index) or a conventional index (MSCI World 1200 Index or S&P Global 1200 Index), (2) a dummy variable indicating if the size of the fund is above the median of the funds in our sample or not and (3) a dummy variable indicating if the total expense ratio of the fund is above the median of the funds in our sample or not. We accept our hypothesis of equality of means when the significance of the T-Student test is above 0.05 (Sarango-Lalangui et al. 2018).

9.3 Data

In order to analyze the financial performance of alternative energy mutual funds, we collected data from the Morningstar and Thomson Reuters EIKON/Datastream database. The Morningstar database identifies 44 alternative energy mutual funds that invest at least 50% of equity assets in worldwide alternative energy firms related to solar power, hydroelectric power, wind power and industrially commercialized nuclear energy in Spain. These data can be accessed free of charge and are available on the Morningstar website. For each of these funds, we collect the ISIN code, monthly total expense ratio, monthly total net assets under management and daily net asset value data expressed in euros from January 2nd, 2007 to July 30th, 2017 from the Thomson Reuters EIKON/Datastream database upon payment of a fee. Additionally, we require mutual funds to have at least 24 months of net asset value data in order to generate reliable estimators (Silva and Cortez 2016; Vidal-García et al. 2016). Our final sample incorporates 42 alternative energy mutual funds that are commercialized in Spain. This sample excludes liquidated and merged funds that could generate survivorship bias. However, there are very few such funds, so we do not believe that this shortcoming will significantly affect our results. We also obtain information relative to several benchmarks such as the MSCI World 1200 Index, 9 Analyzing Alternative Energy Mutual Fund … 207

MSCI World Small Cap Index, MSCI World Large Cap Index, MSCI World Growth Index, MSCI World Value Index, MSCI World Momentum Index, S&P Global 1200 Index and S&P Global Clean Energy Index from the Thomson Reuters EIKON/Datastream database. The euro repo index is provided by NEX Data upon request.

9.4 Results and Implications

Table 9.1 reports annualized performance and risk estimates obtained from Carhart’s four-factor model (Eq. (9.1)) for the equally weighted portfolios of alternative energy mutual funds commercialized in Spain in the 2007–2017 period. Panel A in Table 9.1 presents the average four-factor-adjusted return estimates indicating the number of funds presenting a statistically significant (at the 10% level) positive, not different from zero and negative estimates (No. of +/0/− estimates) by using specialized alternative energy market benchmarks while Panel B reports the alpha estimates considering the conventional global index as the market benchmark. Using the alternative energy index slightly increases the capacity of the model to explain variations in the returns of alternative energy mutual funds as indicated by the coefficient of determination of the regressions (R2). Lower R-squared value indicates that managers are implementing an active management strategy while values of R-squared close to one denotes that portfolio managers are adopting a passive management strategy, that is, they are applying the composition of a market index. The R2 from the model that includes an S&P Global Clean Energy Index (R2 = 0.517) is slightly higher than the R2 of the model incorporating its conventional counterpart, the S&P Global 1200 Index (R2 = 0.494) or the MSCI World 1200 Index (R2 = 0.515). Therefore, the specialized index appears to be more useful for explaining alternative energy returns than conventional indexes, which affects the alpha estimates as shown in Table 9.1. With respect to alternative energy fund performance, we find that the average four-factor-adjusted return from the model including an S&P Global Clean Energy Index is higher than for the model that takes into account the conventional market benchmarks. Concretely, the annualized risk-adjusted return for models introducing a specialized benchmark is 2.8%, on average, while the annualized risk-adjusted return for models with the S&P Global 1200 Index and the MSCI World 1200 Index is −5.1% and −5.3%, respectively, for the 2007–2017 period. This indicates that alternative energy funds earn an average annual return that is 2.8% points higher than the one earned by the S&P Global Clean Energy Index portfolio which includes a diversified mix of clean energy production and clean energy equipment and technology related firms from around the world. On the contrary, the alternative energy funds lose an average annual return that is 5.1 (5.3) percentage points lower than the one earned by S&P Global 1200 Index (MSCI World 1200 Index) which includes a full-diversified portfolio across country and industry. In fact, t-test results for differences in the mean indicate significant 208 C.-P. Martí-Ballester

Table 9.1 Carhart four-factor model financial performance estimates and descriptive statistic Benchmark Alpha Market SMB HML MOM R2 adj Panel A: Alternative energy benchmark S&P Global Clean Energy Index Mean 0.028 0.396 0.006 0.038 0.260 0.517 No. of +/0/− estimates 8/34/0 42/0/0 21/8/13 12/28/2 36/6/0 Standard deviation 0.030 0.107 0.259 0.195 0.156 Max 0.095 0.577 0.385 0.541 0.555 Min −0.038 0.159 −0.414 −0.641 −0.127 Panel B: Conventional benchmarks MSCI World 1200 Index Mean −0.053 0.897 0.222 −0.169 −0.032 0.515 No. of +/0/− estimates 0/28/14 42/0/0 24/18/0 3/17/22 13/20/9 Standard deviation 0.032 0.118 0.275 0.247 0.271 Max 0.004 1.330 0.666 0.234 0.356 Min −0.163 0.745 −0.209 −0.971 −0.930 S&P Global 1200 Index Mean −0.051 0.760 0.216 −0.140 0.067 0.494 No. of +/0/− estimates 0/29/13 42/0/0 23/19/0 4/23/15 25/11/6 Standard deviation 0.032 0.114 0.279 0.238 0.248 Max 0.012 1.082 0.708 0.246 0.410 Min −0.157 0.584 −0.218 −0.938 −0.754 differences at the 95% level in the mean alphas for models using a specialized benchmark and those using the S&P Global 1200 Index (T-test: 11.6843; p-value: 0.000) and MSCI World 1200 Index (T-test: 12.0118; p-value: 0.000). The t-test results for differences in the mean alphas for models using conventional benchmarks are not significant (T-test: −0.3125; p-value: 0.7555) indicating similar alphas using both conventional benchmarks. This is congruent with the findings of Reboredo et al. (2017), who point out that investors in alternative energy mutual funds are paying a premium for adopting alternative energy principles. This could be due to firms that implement sustainable energy practices in their business strategies not improving their long-term corporate financial performance as indicated by Martí-Ballester (2017), leading them to underperform other firms that adopt conventional management strategies. Using the S&P Global Clean Energy Index, 8 of the 42 alternative energy funds significantly exceed their market benchmark while 34 of 42 perform similarly to the market benchmark. Taking into account conventional market benchmarks, we find that 30.95% (13 of the 42 funds)—using the S&P Global 1200 Index as a benchmark—and 33.33% (14 of the 42 funds)—employing the MSCI World 1200 Index as a benchmark—of the alternative energy funds obtain significantly negative performance while none of the alternative energy funds are able to significantly outperform the conventional market benchmarks. The funds reaching a 9 Analyzing Alternative Energy Mutual Fund … 209 risk-adjusted return close to zero oscillate between 69.05% (29 of the 42 funds) for models using the S&P Global 1200 Index and 66.66% (28 of the 42 funds) for models incorporating the MSCI World 1200 Index as a benchmark, respectively. At the individual fund level, these results indicate that investors who transfer their savings into any of the 13–14 alternative energy funds with significantly negative risk-adjusted return are paying a premium for adopting alternative energy principles in relation to MSCI World 1200 Index or S&P Global 1200 Index portfolios. However, those investors, who transfer their savings into any 28–29 alternative energy funds with performance equal to a full-diversified passive portfolio (MSCI World 1200 Index or S&P Global 1200 Index portfolios), contribute to energy sustainability because these funds finance firms’ alternative energy projects which allow firms and governments to diversify energy supply, reduce carbon emissions, and mitigate climate change without sacrificing investors’ long-term financial performance. Different authors, such as Reboredo et al. (2017) and Ferreira et al. (2013), point out that individual investment fund characteristics, such as size and management fees, may affect the financial performance of mutual funds. Table 9.2 reports the effect of mutual fund size on the estimated alphas and risk factors obtained from Carhart’s four-factor model in Eq. (9.1) using the S&P Global Clean Energy Index as a benchmark for the 2007–2017 period. Panel A of Table 9.2 presents the average four-factor-adjusted return estimates for alternative energy funds whose average total net asset value expressed in euros from 2007 to 2017 is above the median while Panel B reports the alpha estimates considering alternative energy funds whose size is below the median. Our evidence indicates that larger alternative energy funds obtain a positive but lower mean risk-adjusted return than that reached by smaller alternative energy

Table 9.2 Carhart four-factor model financial performance estimates by size Benchmark Alpha Market SMB HML MOM R2 adj Panel A: Large Funds S&P Global Clean Energy Index Mean 0.027 0.409 0.017 0.027 0.245 0.539 No. of +/0/− estimates 5/14/0 19/0/0 9/5/6 6/12/1 16/3/0 Standard deviation 0.033 0.117 0.278 0.192 0.170 Max 0.095 0.577 0.385 0.296 0.555 Min −0.034 0.159 −0.414 −0.641 −0.127 Panel B: Small Funds S&P Global Clean Energy Index Mean 0.030 0.387 0.008 0.056 0.295 0.516 No. of +/0/− estimates 3/16/0 19/0/0 10/4/5 5/13/1 17/2/0 Standard deviation 0.028 0.083 0.225 0.212 0.143 Max 0.083 0.570 0.295 0.541 0.507 Min −0.038 0.191 −0.359 −0.424 −0.013 210 C.-P. Martí-Ballester funds. Concretely, the annualized risk-adjusted return for large funds is 2.7%, on average, while the annualized risk-adjusted return for small funds is 3.0% for the 2007–2017 period. Thus, small alternative energy funds earn an average annual return that is 0.3 percentage points higher than the one earned by large alternative energy funds. However, the T-test shows that this difference is not significant (T-test: −1.1517; p-value: 0.2570). This indicates that alternative energy mutual funds do not benefit from the existence of scale economies. At the individual fund level, our results show that 5 of the 19 large alternative energy funds significantly exceed their market benchmark while 14 of 19 perform similarly to the market benchmark. In a similar way, we find that 3 of the 19 small alternative energy funds significantly outperform the S&P Global Clean Energy Index as a market benchmark while 16 of 19 perform similarly to the market benchmark. This indicates that 26.32% of the large alternative energy funds and 15.79% of the small alternative energy funds in our sample are able to beat the market. The existence of economies of scale would have allowed alternative energy mutual funds to reduce their operating costs. Table 9.3 presents the impact of total expense supported by alternative energy funds on their financial performance measured by the Carhart four-factor model in Eq. (9.1) using the S&P Global Clean Energy Index as a benchmark for the 2007–2017 period. Panel A of Table 9.3 reports the average four-factor-adjusted return estimates for alternative energy funds whose average total expense ratio (TER) from 2007 to 2017 is above the median while Panel B presents the alpha estimates considering alternative energy funds whose TER is below the median. Our findings show that alternative energy mutual funds charged with higher TER obtain a positive but lower mean risk-adjusted return than that reached by funds with lower TER. Concretely, the annualized risk-adjusted return for alternative energy mutual funds charged with

Table 9.3 Carhart four-factor model financial performance estimates by total expense ratio Benchmark Alpha Market SMB HML MOM R2 adj Panel A: Higher Total Expense Ratio (TER) S&P Global Clean Energy Index Mean 0.024 0.423 −0.014 0.062 0.254 0.560 No. of +/0/− estimates 4/18/0 21/1/0 10/4/8 7/15/0 19/3/0 Standard deviation 0.028 0.110 0.290 0.108 0.144 Max 0.069 0.577 0.385 0.296 0.555 Min −0.034 0.163 −0.414 −0.220 0.042 Panel A: Lower Total Expense Ratio (TER) S&P Global Clean Energy Index Mean 0.032 0.370 0.027 0.015 0.266 0.475 No. of +/0/− estimates 4/18/0 21/1/0 11/6/5 5/15/2 17/5/0 Standard deviation 0.031 0.096 0.222 0.252 0.168 Max 0.095 0.570 0.295 0.541 0.507 Min −0.038 0.159 −0.359 −0.641 −0.127 9 Analyzing Alternative Energy Mutual Fund … 211 higher TER is 2.4%, on average, while the annualized risk-adjusted return for funds with lower TER is 3.2%, on average, for the 2007–2017 period. However, the T-test (T-test: −0.8159; p-value: 0.4194) indicates that this difference is not significant. This could be due to better managers receiving higher management fees, reducing the funds’ after-fee risk-adjusted return to the level of mutual funds managed by poor managers, which indicates that the alternative energy mutual fund industry is not very competitive. At the individual fund level, our results show that 4 of the 22 funds charged with higher TER and with lower TER significantly exceed their market benchmark. This indicates that 18.18% of the alternative energy funds charged with higher TER is able to beat the market. Managers of these four funds could adopt a differentiation-based strategy focused on offering better management services which are transferring into investors increasing their wealth. The same percentage of funds charged with lower TER is able to outperform the market. Managers of these four funds with lower TER could adopt a cost-based strategy that allows them to outperform the market by charging lower TER. The expenses charged by the other managers in our sample lead 66 mutual funds in our sample to reach a similar after-fee risk-adjusted return to market benchmark.

9.5 Conclusions

This paper provides new evidence of the relationship between financial performance and renewable energy strategy for the alternative energy mutual funds commercialized in Spain from 2007 to 2017 at the individual level. Using a Carhart (1997) four-factor model, we examine and compare the alternative energy funds’ risk-adjusted return obtained from the model that includes the renewable energy global index (S&P Global Clean Energy Index) with those obtained from the models that incorporate a global market index (S&P Global 1200 Index and MSCI World 1200 Index) as a benchmark. While previous researchers have limited their attention to analyzing the financial performance of alternative energy mutual funds at the aggregate fund level, we focus our study on the individual fund level, which enables us to detect negative and positive financial performance that could be canceled at the aggregate fund level. Different and interesting results have emerged from this research. First, our empirical evidence shows that the use of a renewable energy index or conventional indexes affects alpha estimates, the renewable energy index being slightly more powerful for explaining alternative energy fund returns. While 19.05% of alternative energy mutual funds significantly exceed the S&P Global Clean Energy Index, none of them are able to outperform the conventional global indexes used as benchmarks, i.e., S&P Global 1200 Index and MSCI World 1200 Index. None of the alternative energy mutual funds significantly underperform the alternative energy global market benchmark. However, using the conventional global index as a benchmark, we find that around 31–33% of alternative energy mutual funds 212 C.-P. Martí-Ballester significantly underperform the market. This could indicate that firms composing a conventional global index perform better than firms implementing the sustainable energy strategies that make up the renewable energy index. Therefore, investors in alternative energy mutual funds could be paying a premium for decarbonization, which is congruent with modern portfolio theory. Governments could give more grants to firms that adopt sustainable energy activities to compensate for their higher initial investments, which could be reducing the benefits of facilitating the decarbonization of the global economy. Second, our examination of the size and cost effects on alternative energy fund financial performance reveals no significant differences either between the annualized risk-adjusted return of alternative energy funds supporting high total expense ratios and those supporting low total expense ratios or between the annualized risk-adjusted return of large alternative energy funds and those reached by small alternative energy funds. This indicates that (1) there are no economies of scale in the alternative energy fund industry and (2) the existence of a peer effect that may lead better managers to benefit from charging higher fees until mutual funds reach similar financial performance to those managed by their peers. We conclude that alternative energy fund investors could pay a premium for adopting green criteria in their investment-making decisions at the aggregate fund level. However, when we examine the alternative energy fund financial performance at the individual fund level, we find that around 66.66–69.05% of alternative energy mutual funds perform similarly to the market contributing to energy sustainability. Investors in these mutual funds (1) gain some utility from investing according to their green beliefs without sacrificing their long-term financial utility based on maximization of end-of-period wealth and (2) are able to finance firms’ renewable energy projects—by means of mutual funds—which reduce the greenhouse gas emissions and diversify energy supply, favoring the European decarbonization objectives and energy sustainability. This study has one limitation; it considers alternative energy funds commercialized in Spain. Future research should extend the sample to include alternative energy mutual funds commercialized around the world.

References

Bauer R, Koedijk K, Otten R (2005) International evidence on ethical mutual fund performance and investment style. J Bank Finance 29(7):1751–1767 Bollen NP, Busse JA (2001) On the timing ability of mutual fund managers. J Finance 56 (3):1075–1094 Carhart MM (1997) On persistence in mutual fund performance. J Finance 52(1):57–82 Ferreira MA, Keswani A, Miguel AF, Ramos SB (2013) The determinants of mutual fund performance: a cross-country study. Rev Finance 17(2):483–525 Freeman RE (1984) Strategic management: a stakeholder perspective. Pitman, Boston, p 13 Friedman M (1970) A theoretical framework for monetary analysis. J Polit Econ 78(2):193–238 9 Analyzing Alternative Energy Mutual Fund … 213

Ibikunle G, Steffen T (2017) European green mutual fund performance: a comparative analysis with their conventional and black peers. J Bus Ethics 145(2):337–355 Jensen MC (1968) The performance of mutual funds in the period 1945–1964. J Finance 23 (2):389–416 (1968) Malkiel BG (1995) Returns from investing in equity mutual funds 1971 to 1991. J Finance 50 (2):549–572 Markowitz H (1952) Portfolio selection. J Finance 7(1):77–91 Martí-Ballester CP (2017) Sustainable energy systems and company performance: does the implementation of sustainable energy systems improve companies’ ﬁnancial performance? J Clean Prod 162:S35–S50 Newey WK, West KD (1987) Hypothesis testing with efﬁcient method of moments estimation. Int Econ Rev 777–787 Organization for Economic Co-operation and Development (2014) Energy Policies of IEA Countries: European Union 2014 Review. OECD Publishing, France Paetzold F, Busch T (2014) Unleashing the powerful few: sustainable investing behaviour of wealthy private investors. Organ Environ 27(4):347–367 Reboredo JC, Quintela M, Otero LA (2017) Do investors pay a premium for going green? Evidence from alternative energy mutual funds. Renew Sustain Energy Rev 73:512–520 Ruiz-Romero S, Colmenar Santos A, Alonso Gil M (2012) EU plans for renewable energy. An application to the Spanish case. Renew Energy 43:322–330 Sarango-Lalangui P, Álvarez-García J, del Río-Rama MDL (2018) Sustainable practices in small and medium-sized enterprises in Ecuador. Sustainability (2071-1050) 10(6) Silva F, Cortez MC (2016) The performance of US and European green funds in different market conditions. J Clean Prod 135:558–566 Vidal-García J, Vidal M, Boubaker S, Uddin GS (2016) The short-term persistence of international mutual fund performance. Econ Model 52:926–938 Part III Energy Sustainability Technologies Chapter 10 Wind Energy

J. Peuteman

Abstract The present chapter starts with an overview of some basic concepts concerning electrical power generation using wind turbines. Basic topics like the Betz limit and the importance of the power curve are explained. Harvesting energy is mainly a challenge due to the irregular behavior of the wind which explains the development of a large number of different wind turbine types. The dominance of three-bladed horizontal axis wind turbines is explained. The kinetic energy in the wind is used to drive the rotor, i.e., the translational motion of the wind is converted into a rotary movement of the rotor blades. Fluid mechanics are used to explain this conversion. Special attention goes to the drivetrain containing the generator which converts mechanical power into electrical power. The use of asynchronous generators, doubly fed induction generators, and synchronous generators are studied and compared. Finally, a number of technical challenges are considered. These challenges include the impact of the ﬂuctuations of the power generation on the power balance of the electrical grid.

Keywords Betz limit Á Power curve Á HAWT Á Tip speed ratio Absolute and relative wind speed Á Induction generator Á Doubly fed induction generator Á Synchronous generator Á Power balance

10.1 Introduction

The use of wind energy to drive mechanical loads is very old (Gasch and Twele 2002). Mankind harnessed the energy in the wind in order to grind cereals, to pump water, to irrigate farmland, to drive forward sailing ships, to lift heavy loads, and to saw wood. Due to the rise of steam engines, the rise of combustion engines, and

J. Peuteman (&) M-Group (Mechatronics), KU Leuven, Campus Bruges, Spoorwegstraat 12, 8200 Bruges, Belgium e-mail: [email protected]

© Springer Nature Singapore Pte Ltd. 2019 217 E. Motoasca et al. (eds.), Energy Sustainability in Built and Urban Environments, Energy, Environment, and Sustainability, https://doi.org/10.1007/978-981-13-3284-5_10 218 J. Peuteman

finally the ascent of electrical motors, the use of wind energy to drive mechanical loads has decreased dramatically. Experiments to drive an electrical generator using the kinetic energy of the wind date from the 40s and the 50s of the previous century and even from the nineteenth century. Such experiments have been performed in the USA and in Europe. The first mass production of wind power plants has been realized by the German constructor Allgaier (Heier 2006). These wind turbines were used to supply electricity to farms lying far from the public electrical grid. The Allgaier wind turbines contained generators of 6 kW or 10 kW. Due to cheap fossil fuels in the 60s, the importance of wind energy to generate electrical power was very limited. Due to the oil crisis in 1973 and a growing concern for the environment, a new future for electrical power generation based on wind turbines appeared. Due to limitations of fossil fuel resources, in order to limit the exhaust of harmful gases (like CO2) in the atmosphere and to reduce the impact on the climate of our planet, there is a broad consensus to stimulate the use of renewable and sustainable energy. Different renewables contribute to the realization of a sustainable future. Renewables like solar energy, geothermal energy, the use of biomass, tidal energy, hydroelectricity, and wind energy gain importance (Boyle 2012). Research, including subsidy programs, stimulated the development, production, and use of ever-larger numbers of wind turbines across the world. There is a tendency to construct ever-larger wind turbines (e.g., Enercon E-126 7.5 MW turbine, Vestas V164 9.5 MW turbine, and Siemens SG 8.0-167DD 8 MW turbine). There is an evolution to group wind turbines in a so-called wind farm. Onshore wind farms are quite common but also offshore wind farms gain importance (Tavner 2012; Twidell and Gaudiosi 2009). Especially, in densely populated countries, the construction of offshore wind farms provides possibilities to increase the installed power capacity. Although offshore wind farms require large investment costs and high maintenance costs, the higher offshore wind speeds account for a higher power output. Offshore wind farms are far away from the populated areas which imply that the NIMBY syndrome has a lower impact. These efforts imply a cumulative worldwide installed wind power capacity of 539 GW (in 2017) with 52 GW installed in 2017 (according to the Global Wind Energy Council GWEC: http://gwec.net/global-figures/graphs/). When restricting to Europe, 17GW has been installed in 2017 giving a cumulative capacity of 178 GW, i.e., approximately a growth of 10% in a single year. When restricting to North America, 8 GW has been installed in 2017 giving a cumulative capacity of 105 GW. The rise of offshore wind energy is also important. During 2017, 4.3 GW of offshore wind power capacity has been installed implying almost 19 GW of offshore wind power capacity is available worldwide. 10 Wind Energy 219

10.2 Extracting Power from the Wind 10.2.1 The Betz Limit

A wind turbine extracts kinetic energy from the wind. Actually, energy from a translational motion is used to obtain a rotation of the rotor blades. The rotating rotor blades drive the rotor of an electrical generator which converts mechanical power into electrical power. Wind, having a speed v, is an airflow with mass m which contains kinetic energy 0:5mv2. The kinetic energy is proportional with the square of the speed v but when considering an area A (Fig. 10.1), the flowing air mass is proportional with A and the speed v. The flowing air mass is proportional with the air density q. This implies the kinetic power P equals

P ¼ 0:5 qAv3 ð10:1Þ

The power is proportional with the cube of the wind speed v. Variations of the wind speed imply large variations in the available power and the electrical power generated by a wind turbine. In order to have a large power (a lot of kinetic energy), the swept area A of the rotor blades must be large. To construct wind turbines having a higher rated power, sufﬁciently large rotor blades are needed (e.g., https://www.vestas.com and http:// www.mhivestasoffshore.com/: Vestas V164 has a rotor diameter of 164 m) (e.g.,

Fig. 10.1 Extracting power from the wind 220 J. Peuteman https://www.enercon.de: Enercon E-126 has a rotor diameter of 127 m) (e.g., https://www.siemens.com: Siemens SG8.0-167DD has a rotor diameter of 167 m). The air density q depends on the temperature and on the elevation (height). The density increases as the temperature decreases (at sea level: q ffi 1:225 kg=m3 at 15 C and q ffi 1:292 kg=m3 at 0 C). The density decreases at higher elevations (lower atmospheric pressure). Notice, however, that the wind speed generally increases at higher elevations. The relationship between the wind speed and the elevation strongly depends on the roughness of the landscape. The roughness of the landscape determines the friction experienced by the air as it moves over the surface of the earth. Especially, in case of a rough landscape (with buildings, trees, and shrubs), there is a lot of friction and the wind speeds strongly depend on the elevations (Masters 2004). Friction coefficients which allow to calculate the wind speed at different elevations can be found in the literature (Jain 2011). Unfortunately, it is impossible to extract the entire power 0:5qAv3 from the wind. Only a part of the power can be transferred to the rotor blades of the turbine. Suppose the incoming wind (called the upwind) has a speed v as shown in Fig. 10.1 (inspired by Masters 2004). A part of the kinetic energy is transferred to the rotor blades giving a lower downwind speed noted as vd. It is a realistic assumption that the wind speed equals

v þ v v ¼ d ð10:2Þ b 2 when passing the rotor blades. This implies the power extracted by the rotor blades equals ÀÁ ¼ : q 2 À 2 ¼ : q 3 ð : Þ Pb 0 5 Avb v vd 0 5 CP Av 10 3

Here, CP is the rotor efﬁciency. To extract a maximum amount of power from the wind, this rotor efﬁciency must be maximized. A maximum Pb is obtained when the partial derivative of Pb with respect to vd equals zero. It is left as an exercise to show this partial derivative equals zero, when v v ¼ ð10:4Þ d 3

In case vd ¼ v=3, þ ÀÁ v vd 2 2 3 P ¼ 0:5 qA v À v ¼ 0:5 C ; qAv ð10:5Þ b 2 d P Betz

¼ 16 ¼ : % ﬁ : % with CP;Betz 27 59 3 . This maximum ef ciency of 59 3 is called the Betz limit. In reality, this Betz limit is not obtained and it is a challenge for a wind turbine manufacturer to approach this Betz limit as close as possible. 10 Wind Energy 221 10.2.2 The Power Curve of a Wind Turbine

As already mentioned, the upwind speed v is very important. The power in the wind is proportional with the cube of this upwind speed. A typical power curve of a wind turbine is shown in Fig. 10.2. The power curve shows the relationship between the generated power and the upwind speed v. In case the wind speed is too low, i.e., in case the wind speed is lower than the “cut-in wind speed,” no rotation of the rotor blades is obtained and no power will be generated at all. Notice; however, in general, the power losses associated with these low wind speeds are relatively small since the kinetic energy in the wind is limited. As the upwind speed increases, the generated power increases signiﬁcantly. At the “rated wind speed”, the rated power of the wind turbine is obtained. In case the upwind speed is larger than the rated wind speed, the generated power is not allowed to increase. Due to limitations in the mechanical design of the wind turbine and due to limitations in the electrical design of the generator (and the entire installation), the rated power of the wind turbine must not be exceeded. In case the upwind speed becomes larger than the “shutdown wind speed”, the rotor blades are positioned in the so-called feathering position to protect the wind turbine. This implies that the generated power drops to zero. From the point of view of the electrical grid operator, this is an undesired behavior. Indeed, the grid operator must take care that the generated power equals the consumed power. In case of such a sudden decrease of the generated power, other generation units must ramp up their power to generate all consumed power.

Fig. 10.2 Power curve of a wind turbine 222 J. Peuteman

10.3 Wind Turbine Types

Harvesting energy from the wind remains a challenge mainly due to the irregular behavior of the wind. Indeed, technicians face changes in the wind speed (fast and slow wind speed variations, wind gusts), changes in the wind direction, turbulence, etc. Due to this challenge, mankind constructed a large number of wind turbine types all having advantages and disadvantages (Boyle 2012). First, a distinction can be made between a horizontal axis wind turbine (HAWT) and a vertical axis wind turbine (VAWT). Especially, when considering large wind turbines, horizontal axis wind turbines having three rotor blades are dominant. Figure 10.3 visualizes three-bladed offshore wind turbines at the Thornton Bank (C-Power) in the Belgian part of the North Sea. Vertical axis wind turbines are mainly used for smaller powers. The Darrieus and the Savonius turbine types are the most common ones. They both have the advantage that they can use the power from wind coming from all directions, i.e., no yawing mechanism is needed. When constructing larger wind turbines, horizontal axis wind turbines are dominant. A yawing mechanism is needed to be sure the wind blows perpendicular to the plane of rotation of the rotor blades. An important topic when designing horizontal axis wind turbines is the number of rotor blades. The larger the number of rotor blades, the larger the “solidity.” This solidity indicates what fraction of the swept rotor area is covered by material (Boyle 2012). A large solidity implies a large torque in combination with a low speed of rotation which is well suited to drive a mechanical load (e.g., a Western Wheel wind turbine used to irrigate

Fig. 10.3 Horizontal axis wind turbines having three rotor blades 10 Wind Energy 223 farmland). A low solidity implies a smaller torque in combination with a higher speed of rotation which is better suited to drive an electrical generator. Horizontal axis wind turbines having three rotor blades are dominant. In general, the speed of rotation (expressed in revolutions per minute) is small in comparison with the speed of rotation of the generator inside the nacelle. Therefore, a large number of wind turbines contain a gearbox between the rotor blades and the generator. This gearbox increases the speed of rotation of the generator (implying a lower torque). There also exist wind turbines without a gearbox. These “direct-drive wind turbines” have a slowly rotating generator having the same speed of rotation as the rotor blades.

10.3.1 The Rotor Efﬁciency as a Function of the Tip Speed Ratio

The kinetic energy in the wind is a flow. When this energy has not been harvested, it is lost forever. Therefore, it is a common practice to harvest this energy as much as possible, i.e., to use the wind turbine as much as possible and to have a high rotor efficiency CP. The rotor efficiency depends on the type of rotor, the shape of the rotor, the number of rotor blades, and the tip speed ratio. The tip speed ratio is the ratio between the speed of the tip of a rotor blade and the unperturbed upwind speed. As an example, consider a wind turbine having a rotor diameter of 66 m and a speed of rotation which equals 10 rpm. A rotation with 10 rpm gives a slow impression but the tip speed of the rotor blades is high (35 m=s or 125 km=h). In case of an upwind speed of 10 m=s, a tip speed ratio k ¼ 3:5 is obtained with

tip speed k ¼ ð10:6Þ upwind speed

Although the rotor efficiency CP is lower than the Betz limit, it is important to maximize this rotor efficiency. Depending on the solidity, the maximum rotor efficiency is obtained at different tip speed ratios. The larger the solidity, the smaller the optimal k, i.e., the lower the required speed of rotation as shown in Fig. 10.4 (similar with Gasch and Twele 2002). When the tip speed ratio is too low, a lower efficiency is obtained. When having a low solidity and a low speed of rotation, the wind is blowing throughout the swept area of the rotor blades without “touching” these rotor blades. No energy can be given to the blades. A lower efficiency is also obtained when the tip speed ratio is too high. A rotor blade which moves too fast causes turbulences having a negative impact on the next rotor blade. Notice in Fig. 10.4, with an appropriate design, a somewhat higher rotor efficiency can be obtained when having three rotor blades. When using three rotor blades instead of two rotor blades or one single rotor blade, a lower speed of 224 J. Peuteman

Fig. 10.4 Rotor efﬁciencies depending on the tip speed ratio rotation (and tip speed ratio) is appropriate which reduces the noise production. Reducing the noise production is especially important when considering onshore wind turbines. Moreover, blades having a lower speed suffer less from wear. When a rotor blade is passing the tower of the wind turbine, a dip in the torque is obtained (Heier 2006). When having three rotor blades, the relative size of this torque dip is smaller compared to wind turbines having only two or one single rotor blade. Rotor blades require an important investment, i.e., needing only two rotor blades or only one single rotor blade reduces the total investment cost. Especially, in the 70s and 80s of the previous century a lot of two-bladed wind turbines have been constructed. Due to the higher speed of rotation of the rotor blades, the gearbox needs a lower speed ratio when driving the generator. Today, three-bladed wind turbines are dominant also because the lower speed of rotation is more comfortable in the landscape.

10.4 Fluid Mechanics

Consider a horizontal axis wind turbine. Due to the yawing system, the wind speed direction is perpendicular to the plane of rotation of the rotor blades. When converting the translational motion of the wind into a rotary movement of the rotor blades, two types of forces are important: the drag forces and the lift forces (Boyle 2012; Twidell and Gaudiosi 2009). 10 Wind Energy 225

Figure 10.5 shows an airfoil which is the cross section of a rotor blade. The wind is blowing around the rotor blade which accounts for a drag force as shown in Fig. 10.5 (due to an appropriate streamlining, the drag force is minimized). The shapes of the airfoil are different when comparing the top and the bottom. On top, the speed of the air particles is larger implying a lower pressure according to the law of Bernoulli (with h ¼ cte)

v2 p v2 p þ gh þ ¼ cte ! þ ¼ cte ð10:7Þ 2 q 2 q

At the bottom, the speed of the air particles is lower implying a higher pressure. Due to the pressure difference, an upward force, a lift force, is applied to the airfoil. A distinction is needed between the absolute wind speed and the relative wind speed (Jain 2011). The absolute wind speed is the undisturbed wind speed perpendicular to the plane of rotation. In Fig. 10.6, the rotor blade is moving upward and this implies, based on vector addition, a larger relative wind speed w.r.t. the rotor blades. Due to this larger relative wind speed, a larger lift force is obtained. Notice the drag force. The vector sum of the drag force and the lift force equals a ~ resulting force which can be considered as the vector sum of a force F1 and a force ~ ~ F2. The force F1 is the useful force driving the rotor blades. The forces acting on the rotor blades depend on a number of parameters like the wind speed, the rotational speed of the rotor blades, the cross section of the airfoil (the rotor blade), the pitch angle, and the angle of attack. Figure 10.7 shows the pitch angle and the angle of attack. The pitch angle is the angle between the chord line of the airfoil and the plane of rotation. When considering smaller and cheaper wind turbines, the pitch angle can be fixed. But in case of larger wind turbines, the pitch angle can be adjusted (Gasch and Twele 2002; Jain 2011;Lio2018) to the wind speed in order to optimize the airflow and to maximize the rotor efficiency (e.g., between the cut-in wind speed and the rated wind speed). Between the rated wind speed and the shutdown wind speed, the pitch angle is adjusted in order to decrease deliberately the rotor efficiency CP and limit the generated power to the rated power. Especially, when considering larger wind turbines, the rotor has a twist as a function of the radius r. Indeed, the relative airfoil motion in Fig. 10.6 is proportional to the radius changing the direction of the relative wind speed. In order to obtain, e.g., a constant angle of attack, the pitch angle must be adapted to the radius.

Fig. 10.5 Airfoil with drag and lift forces 226 J. Peuteman

Fig. 10.6 Absolute and relative wind speed

Fig. 10.7 Pitch angle and angle of attack of the rotor

10.5 Electrical Generators

The wind, which blows perpendicular to the plane of rotation of the rotor blades, drives these rotor blades and the electrical generator. In case of large wind turbines, the electrical generator is an alternator generating a three-phase AC voltage. Different types of generators can be used (Manwell et al. 2009) but we restrict ourselves to three important types of generators. We study the induction generator with squirrel cage rotor, the doubly fed induction generator (Abad et al. 2011), and the synchronous generator. 10 Wind Energy 227 10.5.1 Wind Turbines Equipped with an Asynchronous Generator

Figure 10.8 shows a wind turbine where the rotor blades drive the rotor of an induction generator (inspired by Wildi 2006). A gearbox is used to increase the speed of rotation since the speed of rotation of the rotor blades is much lower (e.g., 20 rpm) than the speed of rotation of the rotor of the induction generator (e.g., approximately 1000 rpm or 1500 rpm). The stator of the generator injects its power in a three-phase grid and a transformer is used to increase the voltage level. By increasing the voltage level, lower currents are needed which reduces the copper losses in the grid. The stator of the induction generator is connected with a power grid having a frequency of, e.g., f ¼ 50 Hz implying a synchronous speed (expressed in revolutions per minute)

60 f N ¼ ð10:8Þ S p where p is the number of pole pairs. If the rotational speed is lower than the synchronous speed, the induction machine operates as a motor. Figure 10.9 shows the torque–speed characteristic of the induction machine. In order to operate as a generator, the rotational speed must be higher than the synchronous speed. The speed of rotation N is situated between the synchronous speed and the rated speed. This implies that the speed of rotation allows some changes but these changes are

Fig. 10.8 Wind turbine with an asynchronous generator 228 J. Peuteman

Fig. 10.9 Torque–speed characteristic of an induction machine really limited. Therefore, a wind turbine equipped with an induction generator with a squirrel cage rotor is assumed to have a constant speed. The small speed changes are useful to deal with a sudden wind gust. The additional kinetic energy provided by the wind can be stored as kinetic energy in the rotor blades and the entire drive system. Due to the additional kinetic energy, a limited increase in the speed of rotation is obtained. Figure 10.9 not only shows the speed of rotation N but also the slip s. This slip

N À N s ¼ S ð10:9Þ NS

In case of a subsynchronous speed, the slip s [ 0. In case of a supersynchronous speed, the slip s\0. By using an induction generator with an adjustable number of pole pairs, the speed of rotation can be adjusted to the upwind speed. In case of a low wind speed, the rotor efficiency CP is maximized by having a higher number of pole pairs, i.e., a lower rotational speed giving the optimal tip speed ratio k. In case of a high wind speed, the rotor efficiency CP is maximized by having a lower number of pole pairs, i.e., a higher rotational speed. When realizing two pole pair numbers, a dual speed wind turbine is obtained. Figure 10.10 shows the rotor efficiency as a function of the upwind speed with three different rotational speeds (Masters 2004). 10 Wind Energy 229

Fig. 10.10 Rotor efﬁciency depending on wind speed and speed of rotation

10.5.2 Wind Turbine Equipped with a Doubly Fed Induction Generator

Doubly fed induction machines are used in order to obtain a wind turbine with a so-called variable speed. Figure 10.11 shows such a wind turbine (inspired by Wildi 2006). Notice a gearbox is used to give the rotor of the generator a higher speed of rotation than the rotor blades. The stator of the induction machine is connected with a power grid having a frequency of, e.g., f ¼ 50 Hz which ﬁxes the synchronous speed. Notice a frequency converter (and a transformer) between the grid and the rotor of the induction machine. In case the rotor has a slip s, the frequency of the rotor voltages equals fr ¼ jjs f implying a frequency converter is needed to convert f into fr. To connect the frequency converter with the rotor, slip rings and carbon brushes are used. Consider an induction machine and assume the stator consumes an active power PS. When neglecting the power losses in the stator, an electromagnetic power Pd ¼ PS is sent to the rotor. The electrical rotor power PR ¼ sPS (andðÞ 1 À s PS equals the mechanical power). Consider motor mode with the classical situation with rotor resistances (not doubly fed). Since PS [ 0 and s [ 0, a power PR ¼ sPS [ 0 is dissipated in the rotor windings. NoticeðÞ 1 À s PS [ 0 which means the motor is able to drive a mechanical load. Consider generator mode with the classical situation with rotor resistances. Since PS\0 and s\0, a power PR ¼ sPS [ 0 is dissipated in the rotor windings. NoticeðÞ 1 À s PS\0 which means the rotor blades are driving the generator. Indeed, motor mode is obtained when the speed of rotation is lower than the synchronous speed (s [ 0) and generator mode is obtained when the speed of rotation is higher than the synchronous speed (s\0). 230 J. Peuteman

Fig. 10.11 Wind turbine with a doubly fed induction generator

In case of the classical situation with rotor resistances, the power PR ¼ sPS [ 0 is dissipated into heat, i.e., heat losses occur which reduces the efﬁciency. The frequency converter in Fig. 10.11 extracts the power PR [ 0 from the rotor and injects this power into the grid implying PR is not lost. When considering the doubly fed induction machine, it is important that the power PR can be inverted. In case power is extracted from the rotor (PR [ 0), motor mode is obtained at subsynchronous speed and generator mode is obtained at supersynchronous speed. In case power is injected into the rotor, PR \ 0. In such a situation, motoring mode can be obtained above the synchronous speed (PR \ 0, s\0 implying PS [ 0). Generating mode can be obtained below the synchronous speed (PR\0, s [ 0 implying PS\0). Table 10.1 provides an overview. Generator operation can be obtained below and above the synchronous speed implying the wind turbine is able to operate in a large speed range. When the rated power of the frequency converter increases, jjPR can be larger implying a larger speed range (a larger jjs ) is possible. A larger speed range allows to adjust the speed

Table 10.1 Working modes of a doubly fed induction machine Motor operation Generator operation

Subsynchronous speed (N\NS, s [ 0) PR grid ! stator stator ! grid rotor ! grid grid ! rotor (or dissipated)

Supersynchronous speed (N [ NS, s\0) PR grid ! stator stator ! grid grid ! rotor rotor ! grid (or dissipated) 10 Wind Energy 231 of rotation to the upwind speed, i.e., an appropriate tip speed ratio k can be obtained to maximize the rotor efﬁciency. For instance, the speed of rotation of the generator can vary between 70 and 130% of the synchronous speed. The realization of Fig. 10.11 has the disadvantage slip rings and carbon brushes are needed. To face this disadvantage, there exist brushless doubly fed induction machines.

10.5.3 Wind Turbine Equipped with a Synchronous Generator

Figure 10.12 shows a wind turbine equipped with a synchronous machine (inspired by Twidell and Gaudiosi 2009). Notice no gearbox is used between the rotor blades and the rotor of the generator implying a so-called direct-drive system is obtained (Manwell et al. 2009). Such a wind turbine has a variable speed allowing to adjust the speed of rotation to the upwind speed and to optimize the tip speed ratio k. A frequency converter is needed between the stator of the generator and the electrical grid. Indeed, the amplitude and the frequency of the generated sinusoidal voltage are not ﬁxed since they are proportional to the speed of rotation. Notice a rectiﬁer is needed to excite the rotor of the generator using a DC current.

Fig. 10.12 Wind turbine with a synchronous generator 232 J. Peuteman

Since a direct-drive system is used, the rotor of the generator has a low speed of ¼ N rotation (e.g., 20 rpm). In order to generate a normal frequency fgen p 60,withp the number of pole pairs and N the speed of rotation expressed in revolutions per minute, a sufﬁciently large number of pole pairs p is needed. Notice the entire power must be transferred by the frequency converter, whereas the frequency converter only needs to transfer the smaller rotor power in case of the doubly fed induction machine. Notice in Fig. 10.12, the rotor of the generator has slip rings and carbon brushes. Alternatively, a brushless synchronous generator can be used. A brushless synchronous generator is obtained using permanent magnets in the rotor (Wildi 2006) or using a synchronous generator which actually consists of two generators on the same axis in the same housing. Such a brushless generator contains an excitation generator which is excited by a DC current in the stator and where a three-phase voltage is induced in the rotor. The rotor contains semiconductor rectiﬁers allowing to excite the rotor of the main generator. A three-phase AC voltage will be induced in the stator of the main generator.

10.6 The Wind Speed

As already mentioned, the wind speed has a large impact on the kinetic energy available in the wind and the generated electrical power. The wind speeds depend on the location. Before installing a wind turbine, it is useful to measure and record the wind speeds for at least a year at the considered location. Different types of anemometers allow to measure these wind speeds (Jain 2011). Since weather patterns differ from year to year, it is useful to correlate the measured wind speeds with long-term measurement data available from other wind-measuring sites or meteorological stations. There exist maps (Boyle 2012) giving information concerning the mean wind speeds in Europe, the USA, or other countries. In general, offshore wind speeds are higher than onshore wind speeds. In general, closer to the coastline wind speeds are higher than further inland. Wind speeds are time dependent. A distinction can be made between inter-annual speed variations, annual speed variations (depending on the season or the month), diurnal speed variations, and short-term speed variations. When considering measurement data for a full year, a typical histogram as shown in Fig. 10.13 (inspired by Masters 2004) can be obtained. The histogram shows how many hours a year the wind blows at each wind speed. Such a histogram allows to calculate the average wind speed and to calculate the average of the cube of the wind speed. The average of the cube of the wind speed does not equal the cube of the average of the wind speed. The average kinetic energy in the wind is proportional to the average of the cube of the wind speed. 10 Wind Energy 233

Fig. 10.13 Histogram of the wind speed

The information shown in the histogram of Fig. 10.13 is often presented as a probability density function (PDF) which is a continuous function. The probability density function is often a Weibull PDF expressed as k v kÀ1 v k fvðÞ¼ exp À ð10:10Þ c c c where c is the scale parameter and k is the shape parameter. A Rayleigh PDF is obtained in case k ¼ 2. A wind turbine has an impact on the wind. By extracting kinetic energy, a lower downwind speed is obtained. Moreover, a wind turbine creates turbulence. In case wind turbines are installed too close to each other, especially the downwind turbine copes with this wake effect resulting in a lower generated power. Especially when considering wind farms, wind turbines cope with wake effects resulting in a lower generated power. Also when building two wind farms too close to each other, a wind farm copes with the wake effect of the other wind farm implying a reduced generated power.

10.7 The Power Balance of the Electrical Grid

In an electrical grid, it is important to maintain the power balance. More precisely, the generated active power must equal the consumed active power (this condition also applies for the reactive power, but here we restrict ourselves to the active power) which is actually the law of conservation of energy given as 234 J. Peuteman

Pgen ¼ Pcons ð10:11Þ

The consumed active power Pcons is changing with respect to time and tradi- tionally the generated power Pgen is adapted to the changing active power consumption (Boyle 2012; Wildi 2006). An increasing number of wind turbines make it more difficult to respect the required power balance. The generated power of these wind turbines strongly depends on the changing wind speed. Although weather forecasts are useful, it is still impossible to have a correct prediction of the power generated by wind turbines. Since the kinetic energy in the wind is a flow, energy which has not been harvested is lost forever. Therefore, it is a common practice to harvest this energy as much as possible, i.e., to use the wind turbine as much as possible and to maximize the rotor efficiency CP. Unfortunately, this approach implies a variable generated power Pgen;wind. Suppose for simplicity, the other generating units have a controllable power output Pgen;contr and with Pgen ¼ Pgen;wind þ Pgen;contr one obtains that

Pgen;contr ¼ Pcons À Pgen;wind ð10:12Þ

Since the fluctuations in Pgen;wind and Pcons are uncorrelated, large fluctuations in Pgen;contr are generally needed. This implies the need for a large number of, e.g., thermal power plants in standby or operating at partial load in order to deal with these large power fluctuations. Unfortunately, a thermal power plant in standby consumes (fossil) fuels without generating electrical power and the efficiency of a power plant at partial load is low. In order to limit the undesired consumption of (fossil) fuels, other approaches are possible. It can be useful to adapt the consumed power to the generated power of fluctuating renewable energy sources like wind turbines and photovoltaic panels. Actually, demand-side management opens possibilities. From a technical point of view, installing more wind turbines than strictly needed and generating less wind power than possible allow to control the generated wind power. For instance, the Talmudic approach provides an algorithm to obtain a fair restriction of the generated powers in case different generation units belong to different owners (Kim et al. 2011). It can be useful to store the excess of electrical energy (in case wind turbines are producing a large power while the power consumption is low). Afterward, the stored energy can be used when the wind power production is low (e.g., due to low wind speeds) and the power consumption is high. An important way of storing large amounts of electrical energy is the use of pumped hydroelectric energy storage (Boyle 2012). The energy is stored as potential energy by pumping water from a lower elevation reservoir to a higher elevation reservoir. The energy is regained when the water is flowing back from the higher elevation reservoir to the lower elevation reservoir which allows to drive a hydroturbine and an electrical generator. By controlling the flow of the water, the 10 Wind Energy 235 generated power can be adjusted to the needs. Alternatives to the use of pumped hydroelectric storage are the use of compressed air energy storage (CAES), battery storage, energy storage using hydrogen, and others (Xiong et al. 2018).

10.8 Conclusion

The extraction of kinetic energy from the wind in order to generate electrical energy is gaining importance. For instance, when considering the European Union in 2017, onshore wind turbines produced 292 TWh and offshore wind turbines produced 43 TWh of electrical energy. This total of 336 TWh accounts for 11.6% of the total electrical energy production of 2906 TWh (Source: WindEurope—Annual Statistics 2017—https://windeurope.org/). The importance of wind energy is not restricted to the European Union and the growth is a worldwide phenomenon. It is expected that wind will produce more than 20% of the worldwide electricity demand by 2050 (Letcher 2017). Although wind energy is a promising technology, scientists and technicians face a number of important challenges. The present text mentioned the challenges related to the intermittency of the wind (Emeis 2018). The wind speed changes have a major impact on the design of the drive trains of wind turbines. Moreover, important ﬂuctuations in the generated power arise and grid operators remain responsible to maintain the power balance in the electrical grid. Especially, when the number of wind turbines is increasing and the total installed capacity is increasing, then special attention is needed for several topics. Wind turbines need to be installed at challenging locations like urban/suburban areas (Battisti and Ricci 2018) or offshore locations having a large water depth (Cruz and Atcheson 2016). The sector deals with the impact on the wildlife (Köppel 2017), the impact on the power quality issues of the electrical grid (Glasdam 2016), the use of generation units at remote places (requiring the transport of the electrical energy over large distances), the impact of lightning on the installations (van Kuik and Peinke 2016), safety-related aspects, noise limitations, aesthetics, etc. Despite the challenges, the use of wind energy is growing year after year as explained before. For instance, €22.3bn has been invested during 2017 in the European Union (Source: WindEurope—Annual Statistics 2017—https:// windeurope.org/) where €14.8bn has been invested in onshore wind projects and €7.5bn has been invested in offshore wind projects. Increasing industry competition provides investors the opportunity to ﬁnance more wind generating capacity for less money. Especially, due to the drop in the investment costs for offshore wind energy, less money was needed in 2017 to install more capacity. In 2016, €27.5bn was needed to install 0.6 GW (onshore and offshore), whereas €22.3bn was needed to install 1.2 GW in 2017. Tackling technical challenges and the positive economic evolutions suggest a growing future for wind energy. It is expected that wind will produce more than 20% of the worldwide electricity demand by 2050 (Letcher 2017). 236 J. Peuteman

References

Abad G, Lopez J, Rodriguez M, Marroyo L, Iwanski G (2011) Doubly fed induction machine: modeling and control for wind energy generation applications. IEEE Press, Hoboken, New Jersey. ISBN 978-0-470-76865-5 Battisti L, Ricci M (eds) (2018) Wind energy exploitation in urban environment—TurbWind 2017 colloquium. Springer, Cham, Switzerland. ISBN 978-3-319-74944-0 Boyle G (ed) (2012) Renewable energy: power for a sustainable future. Oxford University Press, Oxford. ISBN-978-0-19-954533-9 Cruz J, Atcheson M (eds) (2016) Floating offshore wind energy—the next generation of wind energy. Springer, Switzerland. ISBN 978-3-319-29398-1 Enercon website. https://www.enercon.de. Accessed 12 Aug 2018 Emeis S (2018) Wind energy meteorology: atmospheric physics for wind power generation. Springer, Cham, Switzerland. ISBN 978-3-319-72859-9 Gasch R, Twele J (2002) Wind power plants: fundamentals, design, construction and operation. Solarpraxis, Berlin. ISBN 3-934595-23-5 Glasdam JB (2016) Harmonics in offshore wind power plants: application of power electronic devices in transmission systems. Springer, Cham, Switzerland. ISBN 978-3-319-26476-9 Global Wind Energy Council (GWEC) website. http://gwec.net/global-ﬁgures/graphs/. Accessed 12 Aug 2018 Heier S (2006) Grid integration of wind energy conversion systems. Wiley, West Sussex. ISBN 978-0-470-86899-7 Jain P (2011) Wind energy engineering. Mc Graw Hill, London. ISBN 978-0-07-171477-8 Kim H-M, Kinoshita T, Lim Y (2011) Talmudic approach to load shedding of islanded microgrid operation based on multiagent system. J Electr Eng Technol 6(2):284–292 Köppel J (ed) (2017) Wind energy and wildlife interactions. Presentations from the CWW2015 conference. Springer, Cham, Switzerland. ISBN 978-3-319-51272-3 Letcher TM (ed) (2017) Wind energy engineering: a handbook for onshore and offshore wind turbines. Academic Press, London. ISBN 978-0-12-809451-8 Lio WH (2018) Blade-pitch control for wind turbine load reductions. Springer, Cham, Switzerland. ISBN 978-3-319-75532-8 Masters GM (2004) Renewable and efﬁcient electric power systems. Wiley, Hoboken, New Jersey. ISBN 0-471-28060-7 Manwell JF, McGowan JG, Rogers AL (2009) Wind energy explained: theory, design and applications. Wiley, West Sussex, United Kingdom. ISBN 978-0-470-01500-1 MHI Vestas Offshore Wind website. http://www.mhivestasoffshore.com/. Accessed 12 Aug 2018 Siemens website. https://www.siemens.com. Accessed 12 Aug 2018 Tavner P (2012) Offshore wind turbines: reliability, availability and maintenance. IET, London. ISBN 978-1-84919-229-3 Twidell J, Gaudiosi G (eds) (2009) Offshore wind power. Multi-Science Publishing Co, Essex, United Kingdom. ISBN 978-0906522-639 van Kuik G, Peinke J (eds) (2016) Long-term research challenges in wind energy—a research agenda by the European Academy of Wind Energy. Springer, Cham, Switzerland. ISBN 978-3-319-46919-5 Vestas website. https://www.vestas.com. Accessed 12 Aug 2018 Wildi T (2006) Electrical machines, drives, and power systems. Pearson–Prentice Hall, Upper Saddle River, New Jersey. ISBN 0-13-196918-8 WindEurope website—Annual Statistics 2017. https://windeurope.org/. Accessed 12 Aug 2018 Xiong R, Li H, Zhou X (eds) (2018) Advanced energy storage technologies and their applications. MDPI, Barcelona, Spain. ISBN 978-3-03842-545-8 Chapter 11 Energy Sustainability Through the Use of Thermoelectric Materials in Waste Heat Recovery Systems Recent Developments and Challenges

Emilia Motoasca

Abstract In many applications, only a reduced percentage (industrial processes 30–40%, internal combustion engines ICE 25%, photovoltaic systems PV 15%, etc.) of the primary energy is converted into useful energy. Increased energy efficiency can be realized through better performance of the involved devices, but also through the recovery of the energy losses. Partial recovery of the losses (mainly heat) may be done using mechanical means (turbines, Stirling generators, solar collectors), but still, the heat wasted to environment remains important. In this context, the thermoelectric generators (TEG) based on (novel) thermoelectric materials may offer a good alternative for heat recovery, since TEGs are static devices that in principle do not require maintenance and may work even in harsh environments, like, e.g., space, extreme cold, etc. Besides, TEGs can be used together with PV systems in hybrid installations to harvest more energy from the solar radiation. Up to date, expensive (due to complex manufacturing and scarcity of used materials) and not so efficient (due to low figure of merit ZT < 1 and inefficient MPPT techniques) TEGs have not been applied at large scale for low-grade heat recovery and for (solar) energy harvesting in smart buildings. However, in recent years, many research and development activities around TE-materials are going on worldwide and there is more pressure to increase the energy efficiency of many heat wasting processes and of (solar) energy harvesting in smart buildings. Potential of existent, commercially available TEGs and of emerging TEGs is investigated taking into account their properties, their past and emerging usage related to industrial and residential waste heat recovery and (solar) energy harvesting.

Keywords Thermoelectric materials Á Thermoelectric generator (TEG) Waste heat recovery Á Hybrid thermal/PV systems

E. Motoasca (&) Faculty of Engineering Technology, Department of Electrical Engineering, KU Leuven, TC Ghent, 9000 Ghent, Belgium e-mail: [email protected]

© Springer Nature Singapore Pte Ltd. 2019 237 E. Motoasca et al. (eds.), Energy Sustainability in Built and Urban Environments, Energy, Environment, and Sustainability, https://doi.org/10.1007/978-981-13-3284-5_11 238 E. Motoasca

11.1 Introduction

Increased energy consumption, with more than +30% by 2035, for steady development of industry, transport, agriculture, services, etc., and the quality of life of the citizens will require better management of classical and renewable energy sources, through increased energy efficiency, (Savani et al. 2017; Sathe and Dhoble 2017). In many applications, only a reduced percentage of the primary energy, typical industrial processes 30–40%, internal combustion engines (ICE) 25%, photovoltaic (PV) systems 15%, is converted into useful electrical/mechanical energy. The most efficient generation of electricity from other forms of primary energy converts maximum 35% of the input into electricity, while the rest of the input energy is lost as heat. Combustion of fossil fuel in ICE (internal combustion engines) for generation of electrical/mechanical energy generates also considerable quantities of heat, (Bine Information Service 2016). Increased energy efficiency can be realized through better performance of the involved devices, but also through the recovery of the energy losses. Partial recovery of the losses, mainly heat losses, may be done using mechanical means like turbines, Stirling generators, solar collectors, etc., but still the heat wasted to the environment from industrial and residential processes remains important. Another solution is to convert the heat directly into electricity using the thermoelectric effect, demonstrated already in 1822 by T. J. Seebeck: the application of a temperature difference across a material produces a thermoelectric voltage, the so-called Seebeck effect. Later on, in 1834 J. C. A. Peltier demonstrated that a voltage applied to a thermoelectric material creates a temperature difference between the hot and cold side of the material, the so-called Peltier effect. Thermoelectric generators have been produced already in the 1950s using BiTeSb-alloys, at that time, the highest performance materials for applications close to room temperature, (Nandhakumar et al. 2016; Lewandowski 2015; Haras et al. 2015; Nesarajah and Frey 2016; Wang and Wang 2014). In the 1960s, SiGe alloys were developed by NASA for high-temperature thermoelectric generators in deep space missions and they still operate without any maintenance. Also, the use of thermoelectric devices for cooling is also a widespread application of the thermoelectric materials (Nandhakumar et al. 2016; Wood and Manager 2018; Champier 2017). Waste heat, classified into high-(>650 °C) medium-(230–650 °C) and low-grade (<230 °C) temperature ranges, can be a potential source of electricity even at temperatures below 140 °C. As many recent studies have shown even if a very small proportion of the immense amount of heat losses from industrial and residential processes materials, this will be a step forward in increasing the general energy efficiency. Among other solutions, the heat recovery systems that use thermoelectric materials is a promising alternative for heat losses recovery, (Cheng et al. 2017a; Cowen et al. 2017; Angeline et al. 2017; Panayiotou et al. 2017). During the last decades, significant progress in the development of thermoelectric materials with improved performance in the high-temperature range (>650 °C) has 11 Energy Sustainability Through the Use of Thermoelectric … 239 been made, appropriate to waste heat recovery in high-temperature systems such as automotive exhaust and nuclear energy production systems. However, a significant part of the waste heat at medium and low temperatures from industrial and residential activities and processes is still released in the environment, since the WHR-systems are too expensive or not technically able of recovering medium of low-temperature waste heat. More affordable, cheap and environmental-friendly WHR-systems based on thermoelectric materials are still searched for (Fig. 11.1). Performance, thus efficiency, of any thermoelectric material is directly related to a dimensionless figure of merit zT,

S2r zT ¼ T ð11:1Þ kth where T is the temperature, S is the Seebeck coefﬁcient, r the electrical conductivity and kth the thermal conductivity of the thermoelectric material, respectively (Nandhakumar et al. 2016). Thermoelectric materials with high performance, namely zT > 1.5, should have a high Seebeck coefﬁcient S and a low thermal conductivity kTh (typical feature of nonmetallic materials) and in the same time a high electrical conductivity r (typical feature of metallic materials). These requirements show the challenge in designing high-performance thermoelectric materials since the three characteristic quantities cannot be independently optimized (Lu et al. 2017; Fu et al. 2018).

Fig. 11.1 (Waste) heat sources 240 E. Motoasca

A thermoelectric generator (TEG) is a set of thermoelectric (TE) modules, each composed of tens or hundreds of (n-type, p-type) pairs of connected (in series— electrically and in parallel—thermally) TE-couples, inserted between two heat exchangers, see Fig. 11.2, (Nandhakumar et al. 2016). A TE-module directly converts a part of the thermal energy that passes through them into electricity when working as generator (Seebeck effect). The TE-modules are characterized by a dimensionless figure of merit ZT with a similar form as the ZT-figure of merit of individual (p and n-type) materials in Eq. 11.1, but accounting for the average thermal and electrical properties of the n- and p-types of materials and for construction of the TE-module. The actual ZT-value of a module can be determined experimentally and it is valid only in a narrow temperature range, (Tuley and Simpson 2017). A performant TE-module should have a high figure of merit ZT, thus high electric conductivity, high Seebeck coefficient and low thermal conductivity in a certain temperature range. Also, good thermal and mechanical stability in temperature range of the applications are important. However, a TEG-generator has a nonlinear behavior influenced by numerous parameters. The highest value of ZT reported in research literature is 1.5–2, while common commercially available TE-modules have ZT < 1 at room temperature. The main challenges for the TE-modules in such a TEG-system are the electrical/thermal performance, cost and robustness, (Ren et al. 2017; Assadi et al. 2018). The commercially available TE-materials have optimum performances over a relatively narrow temperature range so that a single material is not suitable for applications in all (low, intermediate and high) temperature ranges, (Ju et al. 2017). The materials needed for commercially available thermoelectric modules are mostly composed of non-abundant (Bi, Te) or of toxic elements (Pb) suitably doped for the required n- and p-type variants. Tellurium (Te) has an abundance of less than 1 ppb and therefore the costs of it fluctuate as the supply is not sufficient. The typical

Fig. 11.2 Thermoelectric (TE) couple and TE-module 11 Energy Sustainability Through the Use of Thermoelectric … 241 commercially available devices, based on Bi2Te3, have the ZT-figure of merit around 1, very low efficiencies at higher temperatures, high degradation at higher temperatures and contain not environmental-friendly materials. All these disadvantages have triggered extended worldwide search for new thermoelectric materials, particularly those comprised of earth-abundant elements, (Tuley and Simpson 2017; Bharti et al. 2018; Cottrill et al. 2018; Laux 2016; Li et al. 2017a). Interest for thermoelectric generators increased in recent years because the extensive research for TE-materials resulted in the discovery of novel possible TE-materials like, chalcogenides, silicides, skutterudites and half-Heuslers, and very recently some organic materials. These (nano)materials have improved figure of merit ZT, make less or no use of scarce elements and have better stability in the needed range of temperatures, (Rogl and Rogl 2017; Schwall and Balke 2018; Lefèvre et al. 2017; Rogl et al. 2018). Some of the abovementioned novel TE-materials will probably be introduced on the market soon, when the remaining manufacturing issues are solved. Some thermoelectric conducting polymers with less stability at temperatures higher than 150 °C reached a ZT of 0.42 at room temperature, but despite the scientific reports confirming the improvement in their thermoelectric properties, the development of efficient polymer based thermoelectric devices is still at initial, lab stage (Sun et al. 2017; Kirihara et al. 2017; Petsagkourakis 2018). It is worth to mention that the TE-materials are still expensive due to reduced market demand and also due to the complex production process. Reported prices vary between 50 and 425 $/kg of thermoelectric material, (Nandhakumar et al. 2016; Champier 2017; Hwang et al. 2017). More demand for TEGs may lower the production costs and the selling prices of the most important TE-materials and modules manufacturers: Alphabet Energy, EVERREDtronics, Evident Thermoelectrics, Ferrotec, GreenTEG, Gentherm, Komatsu Corp., Laird/Nextreme, Hi-Z, TES, Micropelt—Perpetua Power Source Technologies, RedHawk Energy Systems, TECTEG MFR, Tellurex, II-VI Marlow, TEGpower, Chrystal, Yamaha Corp, Thermonmic, ThermoAura. Advantages of TEGs include longer lifespan, no moving parts (no noise), very low operating and maintenance costs (“mount and forget”), reliable operation and use of middle and low-grade thermal energy, no harmful pollutants emitted during their operation and no chemical reactions with environment (Ozone Depletion Potential ODP = 0, Global Warming Potential GWP = 0). Drawbacks that until now limited the extended use of TEGs are the scarcity of the used materials, the (very) complex manufacturing, the quite limited mechanical and thermal stability, their low efficiency and the high prices (Poddar et al. 2017;Jänsch 2017; Siddique et al. 2017). 242 E. Motoasca

11.2 Application of TE-Materials for Waste Heat Recovery (WHR)

The thermoelectric materials/modules market, with a revenue of 279,3 mil$ in 2015 and an estimated revenue of 610 mil$ in 2021, is divided among various types of applications: waste heat recovery (dominant), energy harvesting, direct power generation, and cogeneration, (Wood and Manager 2018). Waste heat recovery (WHR) is the dominant application and has the largest annual growth, but energy harvesting, direct power generation, and cogeneration are also expected to grow at a significant rate during the years to come. Over the years, various types of TEGs have been developed to meet power needs for different automotive, aerospace and defense, industrial and consumer applications (Mustafa et al. 2017;Kütt et al. 2015; Sajid et al. 2017). Thin-film TEGs (µW), micro TEGs (µW-mW), and small TEGs (mW) have been developed to power autonomous sensors and transmitters (Siddique et al. 2017; Gayner and Kar 2016; Yang et al. 2017; Di Paolo Emilio 2017; Iezzi et al. 2017), while large TEGs (>1 W) for industrial applications, automobiles, and solar energy harvesting are still under development (Gayner and Kar 2016). The use of TE-materials (Peltier modules) for cooling applications is known as a reliable method to realize controlled cooling and covers 70–80% of total market TE-materials/modules. TEGs with very good performance have been used for decades in space applications to power devices and sensors in harsh environments. TE-powered sensors have been used for years for monitoring gas pipes in high north, but due to high costs, they remained a niche market despite their advantages above classical batteries: a TEG with 0.01 cm3 volume at 3 K temperature difference can generate 160 µW, that is much more than same volume of a Li-based battery. Existent TEGs have still very low efficiency (around 2–7% efficiency, depending on temperature range), but if the energy recovery is expressed in terms of W or kWh the results are impressive, even if 1% efficiency is accounted: at least 5% savings in terms of fuel (CO2 emissions) using TEGs for heat recovery in ICEs; 1.4 kW TEG-electricity through exhaust heat recovery system of a car (150 kW engine), 5.9 MW of electricity from waste heat recovery system of a 500 MW gas turbine power plant. Improvements in efficiency may be realized not only through new materials but also through improvements in TEG-systems design. In recent years, TE material has been mostly applied for waste heat recovery (WHR) in individual ICEs (for industrial or automotive use) and very little for large-scale industrial applications, despite their considerable advantages over other WHR methods, (Jänsch 2017; Deng et al. 2017): the TEGs are solid-state devices therefore require simple installation and almost no-maintenance (“mount and forget” type of device) and have no significant impact on the overall noise and vibration production of any system. In some cases, the regenerated energy is consumed almost immediately and no energy storage is needed in the system, while in other applications energy storage can be included, (Orr and Akbarzadeh 2017). 11 Energy Sustainability Through the Use of Thermoelectric … 243

A range of 5–70 kW available power (in the form of waste heat) is in average available in the exhaust systems of light duty vehicles. The instantaneous available waste heat depends on the temperature and mass flow rate of the exhaust gas thus implicit on the operating conditions of the engines. The exhaust gases can be directly in contact with a TEG-system that will generate electricity that will further be directly used to support the vehicle electrical system or to contribute to the electricity supply needed for ancillaries (oil and water pumps, sensors, etc.) and (in hybrid vehicles) for vehicle propulsion. Recent studies (Hannan et al. 2017; Demir and Dincer 2017) have shown that the reduction of fuel consumption in passenger cars (in standard tests) could be as high as 20% by using the exhaust waste heat and converting only 10% of it into electricity, (Stobart et al. 2017; Lan et al. 2018). As demonstrated in some research projects, the electric power produced by TEGs using exhaust gases can be used as an extra power source resulting in the general efficiency improvement of the vehicle. However, the careful design of TEGs positioning in the ICE is important, for not interfering in a negative way with the combustion process and with the exhaust after-treatment system (Zhu et al. 2018; Liu et al. 2017). Applications of TEGs even in hypersonic vehicles not only in ground vehicles have been investigated (Cheng et al. 2017b, c; Gong et al. 2018; Cheng et al. 2018). Many new engines and vehicle technologies are being developed in order to make transport more fuel efficient by energy conservation technologies or by energy recovery technologies. Energy conservation technologies such as powertrain electrification and stop-start technologies are already a long-time ongoing, while the energy recovery technologies are currently being increasingly considered: even low efficiency energy recovery can have a significant impact on the overall efficiency due to the large amount of wasted (heat) energy. Energy recovery technologies include regenerative braking, kinetic energy recovery, turbo-generators/ turbo-chargers, and TEGs, (Stobart et al. 2017). Some OEMs have shown interest in applying TEGs to automotives: Nissan Motors tested a TEG SiGe-based prototype capable of producing 35.6 W peak power for a 3 l gasoline engine. One of the other TEG-systems developed by BMW and Ford in cooperation with the US Department of Energy (DoE) was able to generate up to 250 W peak power of electricity under normal driving conditions (50% of the onboard electricity consumption in the BMW Series 5 Sedan). In the HeatReCar project common research of Fiat Thermo-Gen, Fraunhofer, Bosch and Valeo included BiTe, skutterudite and silicide TE-materials, with a 500 W BiTe system demonstrated, while Renault trucks has investigated silicide materials for waste heat recovery within the RENOTER project. The potential of the high-temperature half-Heusler materials in a 1 kW system has been demonstrated by Eberspacher and within the VIPER project, a prototype of a thermoelectric system for automotive applications has been developed by Jaguar Land Rover and 244 E. Motoasca

European Thermodynamics. However, all the abovementioned technologies are still in the research phase, since the cost to generate the power, the system size and efficiency are still major concerns that need to be solved. Also, the performed tests have not revealed yet the long-term thermal and mechanical behavior and failure modes of the TEG-systems, (Stobart et al. 2017). So far, there has not been done enough research on implementing TEGs in industrial settings (e.g., AMETYST project in Germany: Autonomous flexible monitoring units for monitoring technical systems), where most of the processes are highly dynamic and the combination of modeling with in situ measurements is of very recent date. This is an opportunity for evaluating the potential of heat waste recovery using TEGs by answering the following questions: How much power can realistically be extracted from a certain process? What is the optimal design and position and control strategy of the TEGs array for maximizing the harvested power under highly dynamic operation conditions? Photovoltaic (PV) systems extract energy from a reduced part of the solar spectrum, while solar collectors systems discharge 60% of solar heat to the environment. In both types of solar energy harvesting systems, TEGs can be integrated to increase energy efficiency of the systems, (Makki et al. 2016; Omer et al. 2017; Hernandez et al. 2013; Hajji et al. 2017; Wang et al. 2018; Cui et al. 2016, 2017; Contento et al. 2017; El kamouny 2018; Kang et al. 2018). A hybrid photovoltaic-thermoelectric system may generate 160 W/m2 while for an average PV panel max, 140 W/m2 are generated under STC. Other studies (Rehman and Siddiqui 2017) have shown that the portion of the TEG’s electrical power is less than 1% of the total power of the system mainly due to the very low temperature differences between the hot and the col sides of the TEG and it may not be an effective addition to a PV-system, (Hajji et al. 2017; Cui et al. 2017; Chandel and Agarwal 2017). Integration of TEGs with solar energy harvesting requires more attention for the system design and also for the development of specific MPPT (maximum power point tracking) algorithms, (Yusop et al. 2016; Yusop et al. 2017). TEGs can be used also in smart buildings for a new thermal control technology through the integration of TE modules and PV units within the building envelope to actively influence the incoming and outgoing heat flux through the walls and compensate passive heat losses or gains by using solar energy, (Wang et al. 2017; Skovajsa et al. 2017). However, the mentioned technologies are not yet used in actual buildings: so far, only theoretical research and lab experiments around these systems exist. More work should be done for their actual application in buildings and also in new application areas, like, e.g., energy harvesting from pavements/ roads (Guo and Lu 2017; Jiang et al. 2017). 11 Energy Sustainability Through the Use of Thermoelectric … 245

11.3 Estimation and Upscaling of TE-Modules Performance

The maximum thermoelectric efficiency of a TE-module, defined as the ratio between the obtained electrical power on the load resistance and the absorbed heat, can be calculated as pffiffiffiffiffiffiffiffiffiffiffiffiffiffi D þ À ¼ T pffiffiffiffiffiffiffiffiffiffiffiffiffiffi1 ZT 1 ð : Þ g T 11 2 TH 1 þ ZT þ C TH where is the thermoelectric efficiency, ZT is the average dimensionless thermoelectric figure of merit of the TE-module, TC is the cold side temperature, TH is the hot side temperature, and DT is the temperature difference along the thermoelectric sample (DT=TH – TC). Research aiming at the discovery of novel and better thermoelectric materials involves extensive preparation of materials of various compositions/structures and in-depth characterization of their thermoelectric properties. Characterization of thermoelectric materials usually requires determination of the Seebeck coefficient, electrical resistivity and thermal conductivity, or direct measurement of thermoelectric figure of merit of the n-type and p-type variants. Currently, the state-of-the-art facilities for thermoelectric measurement are designed mainly for accuracy and in-depth characterization at expense of the speed. These techniques usually require lengthy procedures to prepare and mount the samples and to conduct measurements in vacuum or inert atmosphere. It is unlikely that these techniques are able to provide sufficient speed required by high-throughput characterization. However, the total performance of a thermoelectric device characterized by the figure of merit ZT depends not only on the individual n-and p-type materials thermal and electrical properties but also on the geometry of the device: the size (height, area of the cross section), the number, and the spacing of the individual legs plus the connectors and the insulation materials, contact materials (Kim et al. 2017; Negash 2017;Lietal.2017b; Zhang et al. 2016; Zhang et al. 2017a, b; Wang et al. 2016; Borcuch et al. 2017; Kempf and Zhang 2016; Gaurav and Pandey 2016), etc. For this purpose, fast, accurate and reproducible methods, and instruments for reliable characterization of thermoelectric modules are needed, (Sun et al. 2018). In Baranowski et al. (2013), a recently developed thermoelectric characterization system of multifunctional probes for high-throughput applications is described. Of all the thermoelectric coefficients, thermal conductivity is one of the most difficult to measure, since in a typical thermal property measurement system, there is always a parasitic heat transfer present, such as conduction along sensor and heater leads or heat exchange between instrument components and the environment. There are many well-established methods for thermal conductivity measurements, but often in the case of thermoelectrics, the thermal conductivity measurement setups are oversimplified and the uncertainties in the measurements underestimated. However, 246 E. Motoasca fast screening of thermoelectric properties may be achieved by direct measurement of ZT using impedance spectroscopy technique. Many different physical models of TEG-generators have been reported in the scientific literature. The reported models (Siouane et al. 2017; Al-Madhhachi and Min 2017; Benday et al. 2017; Wang 2017) are based on electrothermal coupling but are not fully electrical and thus are difficult to use within electrical circuit simulators without considerable design efforts. Most of the studies presented in the literature have evaluated the performance of TEGs at or close to steady-state conditions, while a few works have considered transient aspects. Commercial multiphysics packages, like ANSYS, COMSOL, and Modelica/Dymola do not offer the needed information for accurate evaluation of the dynamic performance of TEGs, (Benday et al. 2017;Högblom and Andersson 2016; Ferreira-Teixeira and Pereira 2018; Meng et al. 2016; Nesarajah and Frey 2017). An effective method for estimating the device level thermophysical properties and cooling performance of micro-TECs capturing has been described in Sun et al. (2018). It has been shown that both boundary and size effects weaken the ZT-figure of merit of the thermoelectric material at the device level. A similar model can be developed for TEGs to help in evaluating the influence of the main parameters on the real figure of merit of the TE-modules. Transient conditions (variable temperature gradients or variable both hot/cold source temperatures) are important because they influence the thermoelectric power generation, (Nesarajah and Frey 2016; Nesarajah and Frey 2017). In energy harvesting applications, a maximum power point tracking method is needed to maximize the power transferred by the TEG source to the load. Use of DC/DC converters as maximum power point trackers (MPPT) is common technology in PV installations in combination with one of the many MPPT algorithms. Just a few of these MPPT algorithms have been applied to TEGs (Maolikul et al. 2017; Chen et al. 2018). In the context of increasing the energy efficiency of industrial processes, thermoelectric generators based on (novel) TE-materials may offer a good alternative for heat recovery, since TEGs are static devices that in principle do not require maintenance and may work even in harsh environments, like, e.g., space, extreme cold, etc. These characteristics make them also suitable for self-powered sensors to monitor for example gas pipes in high north or human parameters. Besides, TEGs can be used together with PV systems in hybrid installations to harvest more energy from the solar radiation, (Omer et al. 2017; Makki et al. 2015). Up to date, expensive (due to complex manufacturing and scarcity of used materials) and not so efficient (due to low figure of merit ZT < 1 and inefficient MPPT techniques) TEGs have not been applied at large scale for low-grade heat recovery in actual industrial processes and in solar energy harvesting. However, they have been used with good results in applications with high-grade waste heat recovery, like space, nuclear energy. 11 Energy Sustainability Through the Use of Thermoelectric … 247

11.4 Design Challenges for Waste Heat Recovery with TE-Materials

A typical TEG is made up of a heat exchange device into which thermoelectric module is connected through to an external circuit, see Fig. 11.3. The design of the heat exchanger (mostly, cooling fins) should ensure the needed cooling of the cold side of the TE-module (Ong et al. 2017). A TEG-array includes a number of TEGs that can be individually switched on/ off and connected in series/parallel by a control unit, see Fig. 11.4. For a practical implementation of smart TEG-arrays in an industrial process where different waste heat flows at various locations, times and temperatures are present, it is necessary to obtain a configuration of TEGs (together with the heat exchangers) arrays having a good/high practical thermal efficiency and a reasonable cost. This configuration should be controlled by a smart MPPT device in order to deliver power at the most favorable operation point, see Fig. 11.5. The performance of any TEG-system depends on the used TE- and other materials, the design of the TE-modules (size, number of legs, interspacing, connectors) and the design parameters of the heat exchangers, (Rana et al. 2017; Tappura 2018). In industrial WHR-applications, the design objective of any TEG system is to maximize the total power output while accounting for the application specific constraints. The constraints include the system installation and maintenance costs, the available space for the installations, the energy storage capabilities, etc. A throughout understanding of the effect of design parameters (TE-modules

Fig. 11.3 Thermo-electric generator—TEG

Fig. 11.4 Smart TEG-array 248 E. Motoasca

Fig. 11.5 TEG-system for a typical industrial process geometry and materials, operating temperature range, etc.) is essential to calculate and optimize the performance of both the module and the integrated TEG-arrays. In the TEG design process, the following design parameters are of prime importance:

• the temperature of the hot side TC and transport properties of the hot side fluid/ gas. • variation in heat source temperature and/or in heat production during the operation of the system • geometry and properties of both TEG and heat exchanger on the cold side. The design process requires an accurate, yet flexible, modeling process based on a simplified electrothermal model of a TEG together with the needed convertors, storage and control units. The initial model will account for the (temporal and spatial distribution of) energy sources, the behavior of the heat exchanger(s), and the characteristics of the external circuits. (Electrical) models for TE-modules exists, but they do not completely describe the behavior of (novel) TE-materials in various, dynamic conditions as the ones common in industrial processes and in solar energy harvesting. As the design proceeds, finer resolution is needed, so mechanical and thermal finite element models will allow the final selection of the design parameters. The estimation of optimal design parameters has been reported for simple, static cases that are far from the varying, dynamic conditions in industrial and residential energy recovery processes, and until now no large-scale industrial TEG-system has been constructed and monitored. However, long-term measurements in real scale 11 Energy Sustainability Through the Use of Thermoelectric … 249 systems may provide important data for the development of optimal parameter design tools. More experience will be needed in future design tools for TEG-systems to build a body of expertise sufficient to be regarded as a design guide. Recent research like (Huang et al. 2017) describes new designs of TEGs for WHR. TEG-systems for automotive applications should use TE-modules with maximum temperatures of 400–500 °C, while in industrial and residential WHR, middle- and low-temperature modules are an important focus. To ensure good performing TEGs stable with time and resistant to oxidation, development of TE-modules with excellent thermal and electrical contacts at the interfaces is critical. Excellent thermal contact must also accommodate fluctuations in thermome- chanical strain between heat exchangers and the TE-modules and variations in module sizes. Mechanical stresses generated by mismatch in the coefficients of thermal expansion and thermal gradients may lead to mechanical failure and performance reduction, (Stobart et al. 2017; Lan et al. 2018).

11.5 Conclusions

The main characteristics and the future prospects for application of TE-materials in waste heat recovery (WHR) have been discussed. Waste heat recovery has a substantial part to play in the improvement of energy efficiency of all types of equipment. All fuel-burning engines reject heat to the environment and have the potential for recovery processes. The value of energy flows may be assessed and their potential judged against the implementation requirements of a recovery process. Also, general industrial processes and HVAC equipment generate large amount of heat to the environment and are therefore good candidates for TE-energy recovery. Thermoelectricity shows considerable potential as an energy recovery process in the form of TEG-systems because of its relative simplicity: TEGs are static devices that in principle do not require maintenance and may work even in harsh environments, like, e.g., space, extreme cold, etc. These characteristics make them also suitable for self-powered sensors to monitor for example gas pipes in high north or human parameters. However, the required TE-materials are still expensive due to the scarcity of materials and complex manufacturing technology. A successful TEG-system design must bring together TE-materials in the form of thermocouples, which in turn are integrated with a heat exchanger to form a complete TEG. Successful implementation will match the design of various TEG-arrays with the application requirements and boundary conditions and will ensure that the TEG continues to deliver an acceptable output power even in dynamic conditions. TEG-systems can be used for energy recovery from exhaust gas in ICEs and also together with PV systems in hybrid installations to harvest more energy from the solar radiation. Up to date, expensive (due to complex manufacturing and scarcity of used materials) and not so efficient (due to low figure of merit ZT < 1 and 250 E. Motoasca inefficient MPPT techniques) TEGs have not been applied at large scale for low-grade heat recovery in actual industrial processes and in solar energy harvesting. However, they have been used with good results in applications with high-grade waste heat recovery, like space, nuclear energy. There are still many challenges to be overcome by the scientists in the coming years for reaching realistic design tools for TEG-systems and for realizing and long-time monitoring of large-scale implementations of affordable and environmental-friendly TEG-systems that will contribute to increased energy efficiency and sustainability.

References

Al-Madhhachi H, Min G (2017) Effective use of thermal energy at both hot and cold side of thermoelectric module for developing efficient thermoelectric water distillation system. Energy Convers Manag 133:14–19 Angeline AA, Jayakumar J, Asirvatham LG, Marshal JJ, Wongwises S (2017) Power generation enhancement with hybrid thermoelectric generator using biomass waste heat energy. Exp Therm Fluid Sci 85:1–12 Assadi MK, Bakhoda S, Saidur R, Hanaei H (2018) Recent progress in perovskite solar cells. Renew Sustain Energy Rev 81:2812–2822 Baranowski LL, Snyder GJ, Toberer ES, Baranowski LL, Snyder GJ, Toberer ES (2013) Effective thermal conductivity in thermoelectric materials. J Appl Phys 204904:1–11 Benday NS, Dryden DM, Kornbluth K, Stroeve P (2017) A temperature-variant method for performance modeling and economic analysis of thermoelectric generators: linking material properties to real-world conditions. Appl Energy 190:764–771 Bharti M, Singh A, Samanta S, Aswal DK (2018) Conductive polymers for thermoelectric power generation. Prog Mater Sci 93:270–310 Bine Information Service (2016) Thermoelectrics: power from waste heat, Themeninfo I, Jan 2017 Borcuch M, Musiał M, Gumuła S, Sztekler K, Wojciechowski K (2017) Analysis of the fins geometry of a hot-side heat exchanger on the performance parameters of a thermoelectric generation system. Appl Therm Eng 127:1355–1363 Champier D (2017) Thermoelectric generators: a review of applications. Energy Convers Manag 140:167–181 Chandel SS, Agarwal T (2017) Review of cooling techniques using phase change materials for enhancing efficiency of photovoltaic power systems. Renew Sustain Energy Rev 73:1342– 1351 Chen WH, Wu PH, Lin YL (2018) Performance optimization of thermoelectric generators designed by multi-objective genetic algorithm. Appl Energy 209:211–223 Cheng F et al (2017a) A thermoelectric generator for scavenging gas-heat: from module optimization to prototype test. Energy 121:545–560 Cheng K, Qin J, Jiang Y, Lv C, Zhang S, Bao W (2017b) Performance assessment of multi-stage thermoelectric generators on hypersonic vehicles at a large temperature difference. Appl Therm Eng 141:456–466 Cheng K, Feng Y, Lv C, Zhang S, Qin J, Bao W (2017c) Performance evaluation of waste heat recovery systems based on semiconductor thermoelectric generators for hypersonic vehicles. Energies 10(4) 11 Energy Sustainability Through the Use of Thermoelectric … 251

Cheng K, Zhang D, Qin J, Zhang S, Bao W (2018) Performance evaluation and comparison of electricity generation systems based on single- and two-stage thermoelectric generator for hypersonic vehicles. Acta Astronaut 151:15–21 Contento G, Lorenzi B, Rizzo A, Narducci D (2017) Efficiency enhancement of a-Si and CZTS solar cells using different thermoelectric hybridization strategies. Energy 131:230–238 Cottrill AL et al (2018) Ultra-high thermal effusivity materials for resonant ambient thermal energy harvesting. Nat Commun 9(1):1–11 Cowen LM, Atoyo J, Carnie MJ, Baran D, Schroeder BC (2017) Review—organic materials for thermoelectric energy generation. ECS J Solid State Sci Technol 6(3):N3080–N3088 Cui T, Xuan Y, Li Q (2016) Design of a novel concentrating photovoltaic-thermoelectric system incorporated with phase change materials. Energy Convers Manag 112:49–60 Cui T, Xuan Y, Yin E, Li Q, Li D (2017) Experimental investigation on potential of a concentrated photovoltaic-thermoelectric system with phase change materials. Energy 122:94–102 Demir ME, Dincer I (2017) Performance assessment of a thermoelectric generator applied to exhaust waste heat recovery. Appl Therm Eng 120:694–707 Deng YD, Hu T, Su CQ, Yuan XH (2017) Fuel economy improvement by utilizing thermoelectric generator in heavy-duty vehicle. J Electron Mater 46(5):3227–3234 Di Paolo Emilio M (2017) Microelectronic circuit design for energy harvesting systems El kamouny K et al (2018) Thermoelectric cooling micro-inverter for PV application. Sol Energy Mater Sol Cells 180:311–321 Ferreira-Teixeira S, Pereira AM (2018) Geometrical optimization of a thermoelectric device: numerical simulations. Energy Convers Manag 169:217–227 Fu Y, Zhang X, Liu H, Tian J, Zhang J (2018) Thermoelectric properties of Ag-doped compound: Mg3-xAgxSb2. J Mater 4(1):75–79 Gaurav K, Pandey SK (2016) Efficiency calculation of thermoelectric generator using temperature dependent material’s properties, May 2016 Gayner C, Kar KK (2016) Recent advances in thermoelectric materials. Prog Mater Sci 83:330– 382 Gong CL, Gou JJ, Hu JX, Gao F (2018) A novel TE-material based thermal protection structure and its performance evaluation for hypersonic flight vehicles. Aerosp Sci Technol 77:458–470 Guo L, Lu Q (2017) Potentials of piezoelectric and thermoelectric technologies for harvesting energy from pavements. Renew Sustain Energy Rev 72:761–773 Hajji M et al (2017) Photovoltaic and thermoelectric indirect coupling for maximum solar energy exploitation. Energy Convers Manag 136:184–191 Hannan MA, Hoque MM, Mohamed A, Ayob A (2017) Review of energy storage systems for electric vehicle applications: issues and challenges. Renew Sustain Energy Rev 69:771–789 Haras M et al (2015) Thermoelectric energy conversion: how good can silicon be? Mater Lett 157:193–196 Hernandez H, Kofuji ST, Van Noije W (2013) Fully integrated boost converter for thermoelectric energy harvesting. In: 2013 IEEE 4th Latin American symposium on circuits and systems, pp 1–3 Högblom O, Andersson R (2016) A simulation framework for prediction of thermoelectric generator system performance. Appl Energy 180:472–482 Huang K, Li B, Yan Y, Li Y, Twaha S, Zhu J (2017) A comprehensive study on a novel concentric cylindrical thermoelectric power generation system. Appl Therm Eng 117:501–510 Hwang J et al (2017) More than half reduction in price per watt of thermoelectric device without increasing the thermoelectric figure of merit of materials. Appl Energy 205:1459–1466 Iezzi B, Ankireddy K, Twiddy J, Losego MD, Jur JS (2017) Printed, metallic thermoelectric generators integrated with pipe insulation for powering wireless sensors. Appl Energy 208:758–765 Jänsch D (2017) Energy and thermal management, air conditioning, waste heat recovery Jiang W et al (2017) Energy harvesting from asphalt pavement using thermoelectric technology. Appl Energy 205:941–950 252 E. Motoasca

Ju C, Dui G, Zheng HH, Xin L (2017) Revisiting the temperature dependence in material properties and performance of thermoelectric materials. Energy 124:249–257 Kang BO, Lee M, Kim Y, Jung J (2018) Economic analysis of a customer-installed energy storage system for both self-saving operation and demand response program participation in South Korea. Renew Sustain Energy Rev 94:69–83 Kempf N, Zhang Y (2016) Design and optimization of automotive thermoelectric generators for maximum fuel efficiency improvement. Energy Convers Manag 121:224–231 Kim TY, Negash A, Cho G (2017) Experimental and numerical study of waste heat recovery characteristics of direct contact thermoelectric generator. Energy Convers Manag 140:273–280 Kirihara K, Wei Q, Mukaida M, Ishida T (2017) Thermoelectric power generation using nonwoven fabric module impregnated with conducting polymer PEDOT:PSS. Synth Met 225:41–48 Kütt L, Millar J, Karttunen A, Lehtonen M, Karppinen M (2015) Thermoelectric applications for energy harvesting in domestic applications and micro-production units. Part I: thermoelectric concepts, domestic boilers and biomass stoves. Renew Sustain Energy Rev 1–26 Lan S, Yang Z, Chen R, Stobart R (2018) A dynamic model for thermoelectric generator applied to vehicle waste heat recovery. Appl Energy 210:327–338 Laux E et al (2016) Development of thermoelectric generator based on ionic liquids for high temperature applications. Eur. Conf. Thermoelectr. 5(4):10195–10202 Lefèvre R, Berthebaud D, Pérez O, Pelloquin D, Boudin S, Gascoin F (2017) Ultra-low thermal conductivity of TlIn5Se8and structure of the new complex chalcogenide Tl0.98In13.12Se16.7Te2.3. J Solid State Chem 250:114–120 Lewandowski CM (2015) Modern theory of thermoelectricity. In: Effect of brief mindfulness intervention on acute pain experience: an examination of individual difference, vol 1 Li C et al (2017a) A simple thermoelectric device based on inorganic/organic composite thin film for energy harvesting. Chem Eng J 320:201–210 Li Y, Wang S, Zhao Y, Lu C (2017b) Experimental study on the influence of porous foam metal filled in the core flow region on the performance of thermoelectric generators. Appl Energy 207:634–642 Liu D, Cai Y, Zhao FY (2017) Optimal design of thermoelectric cooling system integrated heat pipes for electric devices. Energy 128:403–413 Lu W, Xiao R, Yang J, Li H, Zhang W (2017) Data mining-aided materials discovery and optimization. J Mater 3(3):191–201 Makki A, Omer S, Sabir H (2015) Advancements in hybrid photovoltaic systems for enhanced solar cells performance. Renew Sustain Energy Rev 41:658–684 Makki A, Omer S, Su Y, Sabir H (2016) Numerical investigation of heat pipe-based photovoltaic-thermoelectric generator (HP-PV/TEG) hybrid system. Energy Convers Manag 112:274–287 Maolikul S, Kiatgamolchai S, Chavarnakul T (2017) Low-power energy harvesting of thermoelectric battery charger with step-up DC–DC converter: applicable case study for personal electronic gadgets. J Energy Eng 143(4):05017001 Meng JH, Wang XD, Chen WH (2016) Performance investigation and design optimization of a thermoelectric generator applied in automobile exhaust waste heat recovery. Energy Convers Manag 120:71–80 Mustafa KF, Abdullah S, Abdullah MZ, Sopian K (2017) A review of combustion-driven thermoelectric (TE) and thermophotovoltaic (TPV) power systems. Renew Sustain Energy Rev 71:572–584 Nandhakumar I, White NM, Beeby S (2016) Thermoelectric materials and devices Negash A (2017) Direct contact thermoelectric generator (DCTEG): A concept for removing the contact resistance between thermoelectric modules and heat source. Energy Convers Manag 142:20–27 Nesarajah M, Frey G (2016) Thermoelectric power generation: Peltier element versus thermoelectric generator. In: IECON Proceedings (Industrial Electronics Conference), pp 4252–4257 11 Energy Sustainability Through the Use of Thermoelectric … 253

Nesarajah M, Frey G (2017) Optimized design of thermoelectric energy harvesting systems for waste heat recovery from exhaust pipes. Appl Sci 7(6):634 Omer G, Yavuz AH, Ahiska R (2017) Heat pipes thermoelectric solar collectors for energy applications. Int J Hydrog Energy 42(12):8310–8313 Ong KS, Tan CF, Lai KC (2017) Methodological considerations of using thermoelectrics with fin heat sinks for cooling applications. Appl Sci 7(2):1–11 Orr B, Akbarzadeh A (2017) Prospects of waste heat recovery and power generation using thermoelectric generators. Energy Procedia 110:250–255 Panayiotou GP et al (2017) Preliminary assessment of waste heat potential in major European industries. Energy Procedia 123:335–345 Petsagkourakis I et al (2018) Correlating the Seebeck coefficient of thermoelectric polymer thin films to their charge transport mechanism. Org Electron Phys Mater Appl 52:335–341 Poddar VS, Dhokey NB, Garbade RR, Butee SP, Prakash D, Purohit RD (2017) Rapid production of iron disilicide thermoelectric material by hot press sintering route. Mater Sci Semicond Process 71:477–481 Rana S, Orr B, Iqbal A, Ding LC, Akbarzadeh A, Date A (2017) Modelling and optimization of low-temperature waste heat thermoelectric generator system. Energy Procedia 110:196–201 Rehman NU, Siddiqui MA (2017) Performance model and sensitivity analysis for a solar thermoelectric generator. J Electron Mater 46(3):1794–1805 Ren Z, Lan Y, Zhang Q (2017) Advanced thermoelectrics: materials, contacts, devices, and systems. In: Materials Science and Engineering, p 790. CRC Press. https://www.crcpress.com/ Advanced-Thermoelectrics-Materials-Contacts-Devices-and-Systems/Ren-Lan-Zhang/p/book/ 9781498765725. ISBN 9781498765725. CAT# K29106 Rogl G, Rogl P (2017) How nanoparticles can change the figure of merit, ZT, and mechanical properties of skutterudites. Mater Today Phys 3:48–69 Rogl G et al (2018) Nanostructuring as a tool to adjust thermal expansion in high ZT skutterudites. Acta Mater 145:359–368 Sajid M, Hassan I, Rahman A (2017) An overview of cooling of thermoelectric devices. Renew Sustain Energy Rev 78:15–22 Sathe TM, Dhoble AS (2017) A review on recent advancements in photovoltaic thermal techniques. Renew Sustain Energy Rev 76:645–672 Savani I, Waage MH, Børset M, Kjelstrup S, Wilhelmsen Ø (2017) Harnessing thermoelectric power from transient heat sources: Waste heat recovery from silicon production. Energy Convers Manag 138:171–182 Schwall M, Balke B (2018) On the Phase Separation in n-Type thermoelectric half-Heusler materials. Materials (Basel, Switzerland) 11(4):1–17 Siddique ARM, Mahmud S, Van Heyst B (2017) A review of the state of the science on wearable thermoelectric power generators (TEGs) and their existing challenges. Renew Sustain Energy Rev 73:730–744 Siouane S, Jovanović S, Poure P (2017) Equivalent electrical circuits of thermoelectric generators under different operating conditions. Energies 10(3) Skovajsa J, Koláček M, Zálešák M (2017) Phase change material based accumulation panels in combination with renewable energy sources and thermoelectric cooling. Energies 10(2) Stobart R, Wijewardane MA, Yang Z (2017) Comprehensive analysis of thermoelectric generation systems for automotive applications. Appl Therm Eng 112:1433–1444 Sun Y, Xu W, Di C, Zhu D (2017) Metal-organic complexes-towards promising organic thermoelectric materials. Synth Met 225:22–30 Sun D, Shen L, Sun M, Yao Y, Chen H, Jin S (2018) An effective method of evaluating the device-level thermophysical properties and performance of micro-thermoelectric coolers. Appl Energy 219:93–104 Tappura K (2018) A numerical study on the design trade-offs of a thin-film thermoelectric generator for large-area applications. Renew Energy 120:78–87 Tuley R, Simpson K (2017) ZT optimization: an application focus. Materials (Basel) 10(3) 254 E. Motoasca

Wang BL (2017) A finite element computational scheme for transient and nonlinear coupling thermoelectric fields and the associated thermal stresses in thermoelectric materials. Appl Therm Eng 110:136–143 Wang X, Wang ZM (eds) (2014) Nanoscale thermoelectrics, vol 16 Wang BL, Guo YB, Zhang CW (2016) Cracking and thermal shock resistance of a Bi2Te3 based thermoelectric material. Eng Fract Mech 152:1–9 Wang C, Calderón C, Wang YD (2017) An experimental study of a thermoelectric heat exchange module for domestic space heating. Energy Build 145:1–21 Wang P, Wang KF, Wang BL, Cui YJ (2018) Effective thermoelectric conversion properties of thermoelectric composites containing a crack/hole. Compos Struct 191:180–189 Wood L, Manager S (2018) Global thermoelectric energy harvesting market 2018–2028— technologies, devices & applications for thermoelectric generators–Research and Markets, pp 1–4 Yang T, Xie D, Li Z, Zhu H (2017) Recent advances in wearable tactile sensors: materials, sensing mechanisms, and device performance. Mater Sci Eng R Rep 115:1–37 Yusop AM, Mohamed R, Mohamed A (2016) Inverse dynamic analysis type of MPPT control strategy in a thermoelectric-solar hybrid energy harvesting system. Renew Energy 86:682–692 Yusop AM, Mohamed R, Ayob A, Mohamed A (2017) Shapeable maximum-power point-tracking algorithm to improve the stability of the output behavior of a thermoelectric-solar hybrid energy-harvesting system. J Assoc Arab Univ Basic Appl Sci 22:1–8 Zhang H et al (2016) The investigation of thermal properties on multilayer Sb2Te3/Au thermoelectric material system with ultra-thin Au interlayers. Superlattices Microstruct 89:312–318 Zhang Z, Yue H, Chen D, Qin D, Chen Z (2017a) Machine-thermal coupling stresses analysis of the fin-type structural thermoelectric generator. J Electron Mater 46(5):3156–3165 Zhang AB, Wang BL, Wang J, Du JK, Xie C, Jin YA (2017b) Thermodynamics analysis of thermoelectric materials: influence of cracking on efficiency of thermoelectric conversion. Appl Therm Eng 127:1442–1450 Zhu DC, Su CQ, Deng YD, Wang YP, Liu X (2018) The influence of the inner topology of cooling units on the performance of automotive exhaust-based thermoelectric generators. J Electron Mater 47(6):3320–3329 Chapter 12 Optimization Strategy of Sustainable Concentrated Photovoltaic Thermal (CPVT) System for Cooling

Muhammad Burhan, Muhammad Wakil Shahzad and Kim Choon Ng

Abstract Renewable energy resources are susceptible to intermittent power supply, and their standalone operation has prime importance for steady power supply. Solar energy resources have high global availability and potential among all energy sources. Most of areas with high solar energy potential have either dry hot or tropical climate. A major portion of primary energy supply for such area is utilized in their cooling energy needs. In this chapter, a sustainable approach for cooling needs has been proposed using solar energy-based highly efficient concentrated photovoltaic (CPV). A combined cooling system, based upon mechanical vapour compression (MVC), and adsorption chillers have been considered. The MVC chiller utilizes the produced electricity by the third -generation multi-junction solar cells (MJCs). However, adsorption chiller is operated with thermal energy recovered from the cooling of CPV system, which also increases the system efficiency as high as 71%. To handle intermittency, hydrogen production is used primary energy storage system, along with the hot water storage. The complete system configuration is then optimized for standalone operation with optimum components size and minimum cost, using micro-genetic algorithm according to proposed optimization strategy.

Keywords CPVT Á Micro-GA Á Cooling Á Solar efﬁciency Á Hybrid Sustainable

12.1 Background

The sustainability of an energy source depends upon its potential, availability and environmental impact. Currently, fossil fuel-based power systems have potential to meet global energy need with ﬂexible availability, as per demand, without any

M. Burhan (&) Á M. W. Shahzad Á K. C. Ng Water Desalination and Reuse Centre, Biological and Environmental Science & Engineering (BESE), King Abdullah University of Science & Technology, Thuwal 23955-6900, Saudi Arabia e-mail: [email protected]

© Springer Nature Singapore Pte Ltd. 2019 255 E. Motoasca et al. (eds.), Energy Sustainability in Built and Urban Environments, Energy, Environment, and Sustainability, https://doi.org/10.1007/978-981-13-3284-5_12 256 M. Burhan et al. interruption. However, their hazardous emissions are pushing our environment towards irreparable damage. If we look at the renewable energy resources, besides their energy potential, they do not offer steady power supply which is the main requirement of power producing system, i.e. to meet fluctuating demand at any time of the day (IPCC 2012; Shahzad et al. 2018; Ng et al. 2017; Burhan 2015; Burhan et al. 2018a, b). The intermittency and low energy density of renewable energy resources limit their use as primary energy supply (Burhan et al. 2016a, b). On the other hand, application of produced energy is also very important to find the sustainable solution. The energy consumption for cooling application is huge, which can go to 40–50% of the electricity production of some countries especially in desert and tropical regions (Oh et al. 2016). These hot climate regions are also rich in terms of their solar energy potential (Gordon and Viva Books 2008; Shahzad et al. 2017). The only need is to use this solar energy potential effectively and efficiently to fulfil their energy needs (Shahzad et al. 2018a, b). Solar radiations hitting the earth surface have potential higher than the global energy need (Burhan et al. 2017a). However, they are intermittent in nature, with non-uniform availability during diurnal period and zero availability during nocturnal period. To be used as steady power source, sustainable energy storage is needed and hydrogen production provides a long-term and reliable solution for solar energy systems (Saadi et al. 2016). Electrolytic hydrogen production from water can work in a closed loop as the electricity can be produced back in fuel cell with water production, assuming no loss. Therefore, such energy storage configuration can also be used in remote areas without any external supply. On the other hand, conventional electrochemical energy storage in batteries is only suitable for small-scale systems, short-term solution (Burhan et al. 2017b). One of the most important aspects of energy systems is their efficiency and for the solar energy, the system efficiency is very important as it will reduce installation space requirements. For cooling application, mechanical vapour compression (MVC) chillers provide most simple, reliable and efficient solution, by using electricity. However, to produce electricity from solar energy, photovoltaic systems provide most simple configuration for such application. But if we look at the photovoltaic market, it is fully dominated by the single junction-based solar cells which can only respond to certain portion of spectrum, thereby having limitation on their solar energy conversion efficiency (Burhan et al. 2016c, 2018c; Muhammad et al. 2016). Third-generation, multi-junction, solar cell-based concentrated photovoltaic (CPV) system provides highest efficiency among all photovoltaic technology. Up till now, highest efficiency of 46% has been recorded for multi-junction solar cell (Green et al. 2015; Burhan et al. 2018d). Solar-to-hydrogen conversion efficiency of 24% has been reported for CPV under lab conditions and 18% under field conditions (Burhan et al. 2016d), which is almost 2–3-fold higher than only electricity production efficiency of conventional PV (Oh et al. 2015; Burhan et al. 2017c). Despite such potential, the market share of CPV is negligible. In addition, there are very few studies discussing the theoretical modelling and performance potential of concentrated photovoltaic (CPV) system. Even none of the commercial tools, related to optimization and simulation of renewable energy system, i.e. 12 Optimization Strategy of Sustainable Concentrated … 257

INSEL, SOMES, HOMER, RAPSIM, ARES, TRNSYS + HYDROGEMS, SOLSIM (Bernal-Agustin and Dufo-Lopez 2009) and iHOGA, has capability to analyse the concentrated photovoltaic (CPV) system. Therefore, due to greenhouse gas emissions of fossil fuel-based system and the increasing demand of cooling in the hot sunny regions, a sustainable system configuration is proposed as standalone solar system which employs the full potential of received solar energy by using concentrated photovoltaic (CPV) with thermal heat recovery. For steady power supply to the chillers, hydrogen and hot water-based energy storage systems are used to supply electricity and heat according to the system need. As the literature is lacking detailed performance model for CPV system, especially for cooling, therefore, in this chapter, a detailed model and standalone system optimization strategy of CPVT (CPV + Thermal) system is presented for a combined configuration of MVC and adsorption chiller. To counter intermittency of received solar energy, hydrogen and hot water-based energy storage system is proposed whose detailed energy management strategy is also presented. As the system captures the full potential of received solar energy by producing electricity from CPV and revering water heat through solar thermal circuit, therefore, a system efficiency of 71% is expected for solar energy conversion. The system performance model is based upon the temperature and concentration characteristics of multi-junction solar cell. Such combined system is then optimized for the size of each component to have minimum capital cost and uninterrupted output.

12.2 Standalone CPVT-Hydrogen System

Figure 12.1 shows the configuration of sustainable combined cooling system, utilizing concentrated photovoltaic (CPV) as solar energy system. The proposed configuration is operating in standalone operation without any external supply of power and resources (assuming no leak) as it is utilizing combined energy storage system, i.e. hydrogen production and hot water storage. The hydrogen is produced through electrolysis of water using the excess electricity produced by the CPV system, after fulfilling the energy need of the load and system. The produced hydrogen and oxygen are then stored into cylinders to supply back electricity using fuel cell, in case of power deficiency that may happen during cloudy or nocturnal period. The electricity produced by CPV system is supplied to the main DC line through maximum power point tracking (MPPT) devices and DC/DC converter. All of the power-consuming devices, e.g. mechanical vapour compression (MVC) chiller, gas compressor, circulating pumps and solar trackers, are linked to this main DC line through appropriate voltage converter, due to the difference between produced and required voltage. As CPV can only accept beam radiations, therefore, multi-junction solar cells (MJCs) based CPV modules are mounted onto two-axis solar tracking units, for which the power is also supplied from main CPV system. The CPV modules in current study can be of any type either reflector-based 258 M. Burhan et al. units, i.e. Cassegrain assembly or refraction-based units which utilize Fresnel lens, as the performance model discussed in this study is free from concentration assembly type and considers their concentration ratio and optical properties. The main cooling system is based upon the mechanical vapour compression (MVC) chiller for which the electricity is supplied by CPV system and hydrogen-based energy storage unit. CPV units need heat rejection system for their safe operation. As CPV units operate at high concentration, therefore, a large portion of this waste heat is available at high temperature. This heat, instead of rejecting to environment, is recovered and stored in form of hot water storage. This thermal heat storage is then used to operate an auxiliary adsorption cooling unit which not only provides extra cooling from this recovered energy but also increases the overall efficiency of the CPV system up to 71%. In the next section, a detailed performance model for each individual component of the proposed solar-operated sustainable cooling system. The main target for this performance model development is to design a standalone system with uninterrupted supply to the customer needs, irrespective of time and the load requirement but at minimum cost.

Fig. 12.1 Proposed system conﬁguration for sustainable cooling solution using solar energy 12 Optimization Strategy of Sustainable Concentrated … 259

12.3 Performance Model Development

As mentioned in the previous section, the target of proposed system provides uninterrupted supply without any external assistance, utilizing solar energy. Therefore, detailed performance model for each subsystem is developed in this section so that overall system size can be optimized for steady power supply, even during cloudy and nocturnal period, with minimum cost, i.e. avoiding oversized system. Such sustainable cooling system will not only eliminate solar intermittency but it will also work during nocturnal period with maximum system efficiency. Figure 12.2 shows the energy management strategy with which individual components are connecting and interacting with each other. The direct normal irradiance (DNI) or beam radiations intensity data is provided as solar energy input to the system model along with the ambient temperature values as it will affect the overall CPV operating conditions. According to concentrating assembly configuration for CPV modules, i.e. refraction- or reflection-based design and their optical properties, the concentration at the cell area is calculated which then gives the power output of the cell through considered cell characteristics. By analysing the heat loss and the performance of hot water energy recovery circuit, the stored thermal energy is

Fig. 12.2 Energy management order for combined cooling system using CPV-thermal system 260 M. Burhan et al. determined for the operation of adsorption chiller. Depending upon the cooling load requirement, the excess energy produced by the CPV system is determined, which is then converted into hydrogen and oxygen as energy storage. In case of power deﬁciency, stored gases supply back electricity using fuel cell. The performance model for individual component is as follows.

12.3.1 Concentrated Photovoltaic (CPV) System

This subsection is related to the upper part of Fig. 12.2 which gives the electrical power output of the CPV from the beam solar energy input. Based upon the characteristics of the multi-junction solar cell at different cell temperature and concentration values, the total power output of the CPV system can be calculated. In order to simulate the power output of single solar cell, a simple diode model is used which is modiﬁed for the solar cell operating under concentration, as given by Eq. (12.1) (Nishioka et al. 2006). qVC PC ¼ ICVC ¼ VC Io exp À 1 À ISC ð12:1Þ nCkTC

By putting the open-circuit voltage ‘V=VOC’ condition of zero current ‘I=0’ in Eq. (12.1), the diode saturation current factor ‘Io’ factor can be found by Eq. (12.2).

hiISC Io ¼ ð12:2Þ exp qVOC À 1 nC kTC

For the current study, a triple junction solar cell, i.e. InGaP/InGaAs/Ge, is considered for which the concentration and temperature characteristics of open-circuit voltage and short-circuit currents factors are given in Figs. 12.3 and 12.4. If the performance characteristics of solar cell are known at its rated temperature of 25 °C and the corresponding temperature/concentration factors are known, then the values of ‘VOC’ and ‘ISC’ at any operating condition are given by Eqs. (12.3) and (12.4). ð ; Þ¼½ð Þ þ ðÞÀ dVOC ð : Þ VOC TC CC VOC at 25 C C TC 25 12 3 dTC C ð ; Þ¼½ð Þ þ ðÞÀ dISC ð : Þ ISC TC CC ISC at 25 C C TC 25 12 4 dTC C

As the performance of multi-junction solar cell (MJC) is depending upon the concentration at cell area, therefore, based upon received direct normal irradiance 12 Optimization Strategy of Sustainable Concentrated … 261

Fig. 12.3 Open-circuit voltage characteristics for multi-junction InGaP/InGaAs/Ge solar cell

Fig. 12.4 Short-circuit current characteristics for multi-junction InGaP/InGaAs/Ge solar cell

(DNI), area ratio of concentrator ‘Acon’ and solar cell ‘AC’, and the optical efﬁ- ciency of concentrating assembly, the concentration at cell area is given by Eq. (12.5).

¼ Â Acon Â ð : Þ CC Ib gOP 12 5 AC 262 M. Burhan et al.

The optical efficiency of concentrating assembly is depending upon its construction, i.e. single-stage refraction or double-stage reflection. The single-stage refraction is based upon a single Fresnel lens as concentrator, with glass prism rod as homogeniser. However, for double-stage reflection, a pair of optically coated reflectors, i.e. parabolic and hyperbolic, is used in Cassegrain arrangement with glass prism rod as homogeniser. Therefore, Eqs. (12.6a) and (12.6b) provide the optical efficiency calculation for single-stage and double-stage concentrating assemblies, which depends upon, in fact, the optical absorbance and reflectance of respective components.

¼ Â ð : Þ gOP gF gH 12 6a

¼ Â Â ð : Þ gOP gP gS gH 12 6b

As mentioned before, another performance parameter that affects the solar cell performance is its operating temperature. The cell temperature is taken as 40 °C higher than the ambient temperature as the difference between cell and back plate temperature is approximately 10 °C (Yu and Gen 2010) and a temperature difference of 20–30 °C is taken between back plate and the ambient temperature. By gathering the values of all the performance parameter, the net power output by the CPV system can be calculated from Eq. (12.7).

¼ Â Â Â Â Â ð : Þ PCPV gDC=AC gCDC gTr Pmppt NCM NPCPV 12 7

In above equation, the power consumption by CPV system itself is also considered in form of fraction factors which are related to solar tracking unit, DC/DC and DC/AC converters. The voltage converters have certain efficiency, thereby some power is dissipated during voltage conversion. On the other hand, the power consumption by trackers should only be considered during diurnal period, i.e. from sunrise to sunset, which can be calculated from solar geometry model (Burhan et al. 2016e). All of the constant parameters related to the discussed performance models are given in Table 12.1. For Fresnel lens-based single-stage concentration, the optical efficiency of 72–73% (ArzonSolar 2017) is considered as Fresnel lens has transmission efficiency of 90%, due to absorbance of PMMA material of lens, and the receiver efficiency of 92% which is due to the loss at the inlet aperture of the Fresnel lens. For Cassegrain reflectors based double-stage concentration, the optical efficiency of 85% can be considered as silver-coated reflectors have reflective efficiency of 98% (Bennett and Ashley 1965), with further 5 and 1% loss in glass cover plate and homogeniser, respectively. In this chapter, a lower value of 72% is considered as the optical efficiency of the concentrating assembly. The factor ‘Pmppt’ in Eq. (12.7) is the power output from single multi-junction solar cell (MJC) at its maximum power point. As maximum power point tracking (MMPT) device is connected at the output of each CPV unit, therefore, it is 12 Optimization Strategy of Sustainable Concentrated … 263

Table 12.1 Constant parameters for proposed performance model Parameter Value Parameter Value q (C) 1.6021765 Â 10−19 F (As mol−1) 96,485 2 −2 −1 −23 2 −1 −2 k(m kgs K ) 1.3806488 Â 10 t1 (m A ) 1.599 Â 10 2 −1 −1 nC 2t2 (m A °C ) −1.302 2 −1 −2 2 ηmppt 85% t3 (m A °C ) 4.213 Â 10

ηDC/AC 90% n 2

ηCDC 95% Uo (mV) 1065 −1 Urev (V) 1.229 b (mV dec )80 2 −5 −2 r1 (Xm ) 7.331 Â 10 R(Xcm ) 0.438 2 −1 −7 r2 (Xm °C ) −1.107 Â 10 MH2 (g/mol) 2.0159 −1 S1 (V) 1.586 Â 10 CPH (J/kg K) 14,304 −1 −3 S2 (V °C ) 1.378 Â 10 Tcom (K) 306 2 AE (m ) 0.25 ηDC/AC (%) 90 −2 −5 S3 (V °C ) −1.606 Â 10 ηcom (%) 70 assumed the CPV system is always operating at it maximum power point, with maximum possible efficiency at those operating conditions. In order to model this maximum power point tracking device, we know that its main purpose is to operate the solar cell at it maximum power point. Therefore, it depicts that when the solar cell operates its maximum power point, the MPPT device is working. By finding the expression for power output of solar cell at its maximum power point, we can actually model MPPT device, which can be done by equating the first derivative of solar cell power expression, i.e. Eq. (12.1), to zero.

dP C ¼ 0 ð12:8Þ dVC d qVC VCIo exp À 1 À VCISC ¼ 0 ð12:9Þ dVC nCkTC

Equation (12.9) can be further simpliﬁed to ﬁnd the expressions for voltage, current and power produced by the solar cell at its maximum power point, given by Eqs. (12.10), (12.11) and (12.12), respectively. nCkTC qVmppt Vmppt ¼ VOC À ln 1 þ ð12:10Þ q nCkTC qVmppt Imppt ¼ Io exp À 1 À ISC ð12:11Þ nCkTC ¼ Â Â ð : Þ Pmppt gmppt Imppt Vmppt 12 12 264 M. Burhan et al.

12.3.2 Alkaline Electrolyser

This subsection is related to the lower left part of Fig. 12.2 in which the excess power available is supplied to the electrolyser which has certain characteristics. In order to store the excess electricity produced by the CPV system, electrolysers are used to produced hydrogen and oxygen from water electrolysis. The performance and characteristics of electrolyser considered in this study is an alkaline-based electrolyser unit as discussed in (Ulleberg 1998). The power consumption and the production of gases are based upon the voltage–current (IV) characteristics of single electrolyser cell, which is joined in series with similar cell to form a complete unit. Therefore, the main design parameter for electrolyser unit is its power rating or size which depends upon the total number of cells connected in series. As the power characteristics of single electrolyser cell is known, therefore, by knowing the maximum excess power available, the maximum number of required cells for electrolyser units can be calculated by using Eq. (12.13).

ðPMJC;max Â NCM Â NPCPV ÞÀLmin NEC ¼ ð12:13Þ VEC;max Â IEC;max

In the numerator, the expression in bracket deﬁned the maximum rated power of CPV unit, and the complete numerator expression gives the excess power available as ‘Lmin’ shows minimum load requirement. This excess available power is divided by the maximum rated power of single cell of electrolyser, as given in the denominator of Eq. (12.13). As mentioned before, the electrolyser cells are connected in series; therefore, same amount of current is ﬂowing through each cell, given by Eq. (12.14). Â gCDC Pexcess IE ¼ ð12:14Þ NEC Â UE

The voltage characteristics of considered electrolyser, as presented in (Ulleberg 1998), can be found by Eq. (12.15). ! þ t2 þ t3 r þ r T t1 T T2 ¼ þ 1 2 E þð þ þ 2Þ : E E þ UE Urev IE S1 S2TE S3TE log IE 1 AE AE ð12:15Þ

The voltage of electrolyser is essential only to start the electrolysis process. On the other hand, the total production of hydrogen is actually depending upon the amount of current ﬂowing through the electrolyser as it is the amount of electrons that ﬂow and split the water molecules. Therefore, the total output of electrolyser in the form of hydrogen and oxygen gases production is depending upon its current 12 Optimization Strategy of Sustainable Concentrated … 265 consumption, as given by Eq. (12.16). The hydrogen production is twice the oxygen production.

_ ¼ NECIE ¼ _ ð : Þ n gEF 2 n 12 16 E;H2 nF E;O2

The Faraday efﬁciency is assumed to be 95% for a uniform operating temperature of 80 °C.

12.3.3 Proton Exchange Membrane (PEM) Fuel Cell

This subsection is related to the lower right part of Fig. 12.2 when power produced by the CPV system is not enough to meet the load requirement. In order to supply back electricity to the system, a fuel cell-based system is considered which utilizes stored hydrogen and oxygen according to the amount of electricity needed by the system. The fuel cell unit considered for this study is a proton exchange membrane (PEM)-type unit for which the single-cell characteristics are discussed in (Ulleberg 1998). The voltage output of fuel cell is given by Eq. (12.17). IF IF UF ¼ Uo À b : log À R ð12:17Þ AF AF

Like electrolyser, the main design parameter for fuel cell is also its power rating or the amount of cells joined in series. Therefore, maximum number of cell for fuel cell unit is depending upon the maximum amount of power supplied by the fuel cell, i.e. maximum load requirement of the consumer, as given by Eq. (12.18).

¼ Lmax ð : Þ NFC Â Â 12 18 gCDC gDC=AC PFC;max

By knowing the number of cells, the amount of current ﬂowing through the system can be determined by Eq. (12.19) if power deﬁciency is known.

¼ Preq ð : Þ IF Â Â Â 12 19 gCDC gDC=AC NFC UF

Similar to electrolyser, the total output and the rate of gas consumption of fuel cell are depending upon its current production, as given by Eq. (12.20).

_ ¼ NFCIF ¼ _ ð : Þ n gFF 2 n 12 20 F;H2 nF F;O2 266 M. Burhan et al.

12.3.4 Hydrogen Compressor

This subsection is related to the middle right part of Fig. 12.2, under system load. For standalone operation and steady power supply of the proposed solar system, the excess produced energy is converted into hydrogen and oxygen by electrolysis of water. Such produced gases also require a reliable and compact storage system which can readily supply back these gases when needed, without any external assistance. Mechanical gas compression system with cylinder storage provides reliable and long-term gas storage solution with compact size. Therefore, in order to model such compression unit, a thermodynamic performance of mechanical compressor is considered as given by Eq. (12.21) (Li et al. 2009). It is polytropic process equation which is applied with ﬁrst law of thermodynamics in terms of power required to the change the temperature of gas, i.e. before and after compression. () ðÞrÀ1 M T P r P ¼ n_ Â H2 Â CP Â com ta À1 ð12:21Þ com ; H Â E H2 1000 gDC=AC gcom PE

12.3.5 Hydrogen Storage Cylinder

This subsection is related to the lower mid-part of Fig. 12.2. As the compressed gases are stored into cylinders, therefore, it is also very important to model the state of the storage cylinder in form of its pressure, in order to estimate the available energy stored. In order to model the performance of the storage cylinder, an ideal gas Eq. (12.22) considered with compressibility factor ‘Z’ which the stage of the cylinder in form of its pressure as volume is constant. In this study, the value of compressibility factor is obtained from the data table of real hydrogen gas for 3.34 m3 tank at 33 °C.

ntaRTta Pta ¼ Â ZH ð12:22Þ Vta

12.3.6 Hot Water Storage Tank

This subsection is also related to the lower mid-part of Fig. 12.2. The waste heat recovered from the back plate of multi-junction solar cells in CPV system, through hot water circulation, is stored into hot water storage tanks as shown in Fig. 12.1. 12 Optimization Strategy of Sustainable Concentrated … 267

A one-dimensional transient model is developed for stratified hot water tank. The tank is assumed to be divided into finite number of layers with thermal uniformity. An energy balance is applied to control volume of each thermal layer and based upon the mass balance for convective and conductive flows, a set of temperature differential equations are written below to predict the thermal state of hot water storage tanks during operation. ÂÃÀÁÂÃÀÁ+ À þ À Fc;i mcCp;w Tc;o TT;1 FD;j mT Cp;w TT;2 TT;1 Tt À TtÀ1 ¼ T;2 T;1 T ; À T ; VLqwCp;w þ kA T 2 T 1 Dt Dh ð12:23aÞ ÂÃÀÁÂÃÀÁ+ F ; m C ; T ; À À T ; þ F ; m C ; T ; þ À T ; c i c p w T n 1 T n D j T p w T n 1 T n T ; À À T ; T ; À T ; þ þ kA T n 1 T n À T n T n 1 Dh Dh t À tÀ1 TT;n TT;n ¼ V q C ; ð12:23bÞ L w p w Dt ÂÃÀÁÂÃÀÁ+ F ; m C ; T ; À T ; þ þ F ; m C ; T ; þ À T ; þ c i c p w T n T n 1 D j Tp w T n 2 T n 1 ÂÃÀÁ TT;n À TT;n þ 1 TT;n þ 1 À TT;n þ 2 þ F ; m C ; T À T ; þ þ kA À DM j m p w R T n 1 Dh Dh t À tÀ1 TT;n þ 1 TT;n þ 1 ¼ V q C ; L w p w Dt ð12:23cÞ ÂÃÀÁÂÃÀÁ+ F ; m C ; T ; þ À T ; þ þ F ; m C ; T À T ; þ c i c p w T n 1 T n 2 D j T p w R T n 2 ÂÃÀÁ TT;n þ 1 À TT;n þ 2 þ F ; m C ; T À T ; þ þ kA DM j m p w R T n 2 Dh t À tÀ1 TT;n þ 2 TT;n þ 2 ¼ V q C ; ð12:23dÞ L w p w Dt where ‘FC’, ‘FD’ and ‘FDM’ represent the control function for collector discharge, hot water discharge and mixing water discharge, respectively, for which the value can be either 0 or 1 depending upon the state of operation of tank. However, subscripts ‘i’ and ‘j’ represent the tank number associated with the collector discharge, respectively.

12.3.7 Adsorption Chiller

This subsection is related to the middle part of Fig. 12.2, to calculate the thermal energy consumption for cooling. The adsorption chiller is considered as the 268 M. Burhan et al. auxiliary cooling system which works on the heat energy recovered during operation of CPV system. The performance model of the adsorption system is based upon its thermal energy consumption which can be found by knowing the COP of the considered chiller.

CEAD COPAD ¼ ð12:24Þ QAD

The cooling effect of adsorption chiller ‘CEAD’ is an optimization parameter which has to be found out as per available recovered energy.

Lcool Â FRAD QAD ¼ ð12:25Þ COPAD

This heat energy is supplied to the adsorption chiller in form hot water, as given by Eq. (12.26).

: ÀÁ QAD ¼ m ÂCp;w Â TAD À TAD;R ð12:26Þ AD

By knowing the thermal energy requirement of the adsorption chiller, the return hot water temperature can be found by Eq. (12.27). 2 3 Â 4 Lcool FRAD 5 TAD;R ¼ TAD À : ð12:27Þ COPAD Â m ÂCp;w AD

12.3.8 Mechanical Vapour Compression (MVC) Chiller

This subsection is also related to the middle part of Fig. 12.2, but to calculate the electrical energy consumption for cooling. The performance of mechanical vapour compression (MVC) chiller is also depending upon the COP of the considered system. However, the most important factor in the operation of MVC is its operating capacity that is depending upon the hot water storage system. If the supply temperature of hot water storage tank is enough to operate adsorption chiller, then MVC operates in part load conditions according to the share of cooling load taken by adsorption chiller. If the hot water storage unit does not have enough energy to operate adsorption chiller, then MVC operates its full capacity to cover 100% cooling load requirements, as given by Eqs. (12.28) and (12.29). 12 Optimization Strategy of Sustainable Concentrated … 269 ÀÁ if T ; [ T T 1 AD ð : Þ Lcool Â ðÞ1 À FRAD 12 28 LElect ¼ COPCH ÀÁ if T ; \T T 1 AD ð : Þ Lcool 12 29 LElect ¼ COPCH

12.4 Optimization Objective Functions and Strategy

The main objective of current chapter is to develop a sustainable cooling system based upon solar energy which can operate in standalone mode, without any external assistance and supply. Energy storage systems are coupled with primary supply unit, to handle intermittency. Moreover, the performance models for every component, at any operating condition, are developed in the previous section. However, the main question is, what should be the optimum size of each component to meet the mentioned targets. To be precise, the above-mentioned targets have been deﬁned in mathematical form as three objective functions, Eqs. (12.30), (12.31)and(12.32), for the optimization of proposed system conﬁguration. X PSFT ¼ tPF ¼ 0 ð12:30Þ year

L1\STH2ðf Þ À STH2ðiÞ\L2 ð12:31Þ

CAT ¼ CCPV þ CEL þ CFC þ CSTH2 þ CSTO2 þ Ccom þ CAD þ CMVC þ CHWT þ CHX ð12:32Þ

CCPV ¼ðNPCPV Â NCM Â PMJC;maxÞÂ½CCCPV þ ðÞOMCCPV Â CRF ð12:32aÞ

¼ð Â Þ CE=FC NÂÃEC=FC PE=ÀÁFC;max ÀÁ Â CCE=FC þ RCE=FC Â SPPW þ OMCE=FC Â CRF ð12:32bÞ ÂÃÀÁ CSTH2=O2=HWT ¼ STMH2=O2=HWT Â CCSTH2=O2=HWT þ OMCSTH2=O2=HWT Â CRF ð12:32cÞ

¼ Ccom=AD=MVC=HX PcomÂÃ=AD=MVC=HX ÀÁ Â CCcom=AD=MVC=HX þ OMCcom=AD=MVC=HX Â CRF ð12:32dÞ 270 M. Burhan et al. "# i Âð1 þ iÞL CRF ¼ ð12:32eÞ ð1 þ iÞL À 1

1 SPPW ¼ ð12:32fÞ ð1 þ iÞy

The first and most important objective function, Eq. (12.30), is the uninterrupted power supply, which is defined in form of power supply failure time and for uninterrupted power, its value must be zero. Therefore, any configuration for which PSFT factor is greater than zero is not acceptable. In addition, the system is utilizing energy storage system to avoid intermittency. Therefore, the state of stored energy must be maintained within certain limit, given by Eq. (12.31), for steady and optimum operation. It must be noted that only hydrogen storage is considered as water and oxygen storage are directly linked to it. Water and oxygen storage are of the same size as compared to each other but half the size of hydrogen storage, in terms of moles. On the other hand, the hot water storage is taken as variable parameter which has to be optimized according to available thermal energy. Last and most important objective function is the overall cost of the system, given by Eq. (12.32). As the steady power supply can be easily obtained by oversizing system, therefore, the objective of this study is to find out optimum and economical system configuration by meeting the other two objective functions. The overall cost function includes installation/capital cost (CC), operation and maintenance cost (OMC) and replacement cost (RC). The detailed costing functions for each of the system components are given by Eqs. (12.32a)–(12.32d). The factors ‘CRF’ and ‘SPPW’ in Eqs. (12.32e)–(12.32f) represent capital recovery factor and single payment present worth, respectively. To achieve the optimum configuration of proposed system, as per defined objective functions, an optimization strategy is developed in Fig. 12.5, according to energy management strategy shown in Fig. 12.2. However, the shown optimization strategy is implemented by using micro-genetic algorithm (micro-GA) using FORTRAN programming language. The optimization programmes are developed in two parts. The first part deals with the performance simulation of the proposed system configuration, at any operation condition, according to their energy management strategy. The second part deals with the optimization of system sizing parameters according to defined objective function. The optimized parameters considered for current study are normal of CPV modules, number of hot water storage tanks, percentage share of adsorption chiller and initial amount of hydrogen storage. However, other parameters are calculated from the optimization parameters. The micro-GA algorithm is implemented with population size of 5 with maximum 300 generation. The optimization curve and the system performance characteristics are discussed in the next section. 12 Optimization Strategy of Sustainable Concentrated … 271

Fig. 12.5 Optimization strategy for proposed system conﬁguration using micro-GA

12.5 Optimization Results and Performance Analysis

According to defined optimization strategy and objective function, the micro-genetic algorithm (micro-GA) is implemented for the standalone operation of proposed system configuration as per defined energy management strategy and the optimization results are shown in Fig. 12.6. From the results, first of all, it is clear that the 300 generations were enough to obtain the optimum results. However, convergence of the programme, i.e. minimum cost, was obtained after 120 generations. On the other hand, it can also be seen that for all of the configurations, the 272 M. Burhan et al.

PSFT factor is zero, which ensure the steady and uninterrupted power supply to the load. From energy storage point of view, the difference in the state of hydrogen cylinder, before and after the operation, is within the assumed limits of ‘L1 = −10 kg’ and ‘L2 = 35 kg’. Such developed strategy and configuration can be applied to the cooling system of any capacity, to achieve their optimum operating configuration, by utilizing the maximum potential of solar energy in form of CPV system. As mentioned before, CPV system provides the best of the solar energy system with highest high-quality energy conversion efficiency. Based upon the proposed configuration, the performance of the proposed CPV system, as per actual received DNI data, is shown in Fig. 12.7. It can be seen that the CPV system showed maximum electrical efficiency of 27–28%. However, the CPV-Hydrogen system showed maximum efficiency of 18% which is about twofold higher than the efficiency of conventional PV systems. The performance curves highlight the variation in system output during whole day operation. It can be seen that both electrical and hydrogen efficiencies of the CPV system are first increasing and then decreasing towards middle of the day. In case of hydrogen production, the decrease in overall efficiency is due to decline in the electrolyser efficiency. As the intensity of solar radiations increases from morning to noon, the power output of CPV system also increases, which is supplied to electrolyser. With increase in supplied power, its operating voltage also increases, due to which its efficiency declines in inverse manners, as shown in Fig. 12.7. On the other hand, in case of CPV electrical efficiency, the decline in efficiency is due to increase cell temperature (Bennett and Ashley 1965), as discussed before. As the received DNI increases, the concentration at cell area also increases and so as the waste heat. Therefore, electrical efficiency of CPV slightly decreases. But overall electrical and thermal efficiencies of CPV-Thermal (CPVT) remain steady as where electrical efficiency decreases due to increased heat, at the same time, thermal efficiency increases with better heat recover. The maximum value of 71% is presented as solar energy efficiency in case of CPVT system.

Fig. 12.6 System optimization curve for micro-GA 12 Optimization Strategy of Sustainable Concentrated … 273

Fig. 12.7 Performance characteristics of proposed concentrated photovoltaic (CPV) system

The most interesting aspect here is the overall highest efﬁciency of CPVT system. Although all of the converted solar energy is not available in high-grade form, it can be fully utilized as per application, like proposed CPVT-cooling system. The recovered heat is enough to operate adsorption chiller which provides auxiliary support to main mechanical chiller. Thus, the proposed conﬁguration CPVT system provides a sustainable solution to our cooling needs, using combined chiller system, operated by solar energy. However, the hydrogen energy storage provides a reliable and long-term energy storage solution to handle the solar intermittency issues as it can be scaled up to any size of the system and can be stored for long period, as compared to battery, assuming no leak.

12.6 Summary of Chapter

In order to fulfil increasing global cooling needs and due to damaging effects of fossil fuel use, a sustainable solution is proposed using highly efficient solar photovoltaic system, i.e. concentrated photovoltaic (CPV). Due to its operation at high solar concentration, a heat recovery system is also coupled which increases the overall solar energy conversion efficiency of CPVT system as high as 71%. As both electrical and thermal energies are received as output, therefore, a combined cooling system is proposed, which is based upon electricity operated mechanical vapour compression (MVC) chiller and thermal operated adsorption. By keeping in the issue of solar intermittency and need of steady power supply, hydrogen production is utilized as long-term electrical energy storage system, and hot water is considered for thermal energy storage. 274 M. Burhan et al.

The main challenge in such proposed conﬁguration is its optimized design in which all of the system components are of perfect size, according to their need, with minimum cost. Therefore, a detailed performance model and optimization strategy are proposed for standalone operation of CPVT-based combined cooling system. The optimization algorithm and strategy is proposed using micro-genetic algorithm (micro-GA) which ﬁnds out the optimum size of every component of the system, for uninterrupted power supply with minimum system cost. As none of the commercial simulation software is capable of handling CPV, therefore, the proposed strategy can also be integrated with them to enhance their capability. Moreover, such model can be used to optimize CPV system for any application and capacity for steady power supply.

References

ArzonSolar (2017) http://www.arzonsolar.com/wp-content/uploads/2015/02/uModule-Datasheet. pdf. Date retrieved 25 Feb 2017 Bennett JM, Ashley EJ (1965) Infrared reflectance and emittance of silver and gold evaporated in ultrahigh vacuum. Appl Opt 4(2):221–224 Bernal-Agustin JL, Dufo-Lopez R (2009) Simulation and optimization of stand-alone hybrid renewable energy systems. Renew Sustain Energy Rev 13:2111–2118 Burhan M (2015) Theoretical and experimental study of concentrated photovoltaic (CPV) system with hydrogen production as energy storage (Doctoral dissertation) Burhan M, Chua KJE, Ng KC (2016a) Simulation and development of a multi-leg homogeniser concentrating assembly for concentrated photovoltaic (CPV) system with electrical rating analysis. Energy Convers Manag 116:58–71 Burhan M, Chua KJE, Ng KC (2016b) Sunlight to hydrogen conversion: design optimization and energy management of concentrated photovoltaic (CPV-Hydrogen) system using micro genetic algorithm. Energy 99:115–128 Burhan M, Chua KJE, Ng KC (2016c) Electrical rating of concentrated photovoltaic (CPV) systems: long-term performance analysis and comparison to conventional PV systems. Int J Technol 7(2):189–196. https://doi.org/10.14716/ijtech.v7i2.2983 Burhan M, Chua KJE, Ng KC (2016d) Long term hydrogen production potential of concentrated photovoltaic (CPV) system in tropical weather of Singapore. Int J Hydrog Energy 41 (38):16729–16742 Burhan M, Oh SJ, Chua KJE, Ng KC (2016e) Double lens collimator solar feedback sensor and master slave configuration: development of compact and low cost two axis solar tracking system for CPV applications. Sol Energy 137:352–363 Burhan M, Shahzad MW, Ng KC (2017a) Development of performance model and optimization strategy for standalone operation of CPV-hydrogen system utilizing multi-junction solar cell. Int J Hydrog Energy 42(43):26789–26803 Burhan M, Oh SJ, Chua KJ, Ng KC (2017b) Solar to hydrogen: compact and cost effective CPV field for rooftop operation and hydrogen production. Appl Energy 194:255–266 Burhan M, Shahzad MW, Ng KC (2017c) Long-term performance potential of concentrated photovoltaic (CPV) systems. Energy Convers Manag 148:90–99 Burhan M, Shahzad MW, Ng KC (2018a) Sustainable cooling with hybrid concentrated photovoltaic thermal (CPVT) system and hydrogen energy storage. Int J Comput Phys Series 1 (2):40–51 12 Optimization Strategy of Sustainable Concentrated … 275

Burhan M, Shahzad MW, Ng KC (2018b) Energy distribution function based universal adsorption isotherm model for all types of isotherm. Int J Low-Carbon Technol. https://doi.org/10.1093/ ijlct/cty031 Burhan M, Shahzad MW, Oh SJ, Ng KC (2018c) A pathway for sustainable conversion of sun-light to hydrogen using proposed compact CPV system. Energy Convers Manag 165:102– 112 Burhan M, Shahzad MW, Choon NK (2018d) Hydrogen at the rooftop: compact CPV-hydrogen system to convert sunlight to hydrogen. Appl Therm Eng 132:154–164 Gordon JM, Ng KC (2008) Cool thermodynamics. Viva Books Green MA, Emery K, Hishikawa Y, Warta W, Dunlop ED (2015) Solar cell efficiency tables (Version 45). Prog Photovolt Res Appl 23(1):1–9 HOMER (Hybrid Optimization of Multiple Energy Resources). http://www.homerenergy.com/ software.html iHOGA (Improved Hybrid Optimization by Genetic Algorithms). http://personal.unizar.es/rdufo/ index.php?option=com_content&view=article&id=2&Itemid=104&lang=en IPCC (2012) Renewable energy sources and climate change mitigation. Special report of the intergovernmental panel on climate change Li CH, Zhu XJ, Cao GY, Sui S, Hu MR (2009) Dynamic modeling and sizing optimization of stand-alone photovoltaic power systems using hybrid energy storage technology. Renew Energy 34(3):815–826 Muhammad B, Seung JO, Ng KC, Chun W (2016) Experimental investigation of multijunction solar cell using two axis solar tracker. Appl Mech Mater 819:536–540. https://doi.org/10.4028/ www.scientific.net/AMM.819.536 Ng KC, Burhan M, Shahzad MW, Ismail AB (2017) A universal isotherm model to capture adsorption uptake and energy distribution of porous heterogeneous surface. Sci Rep 7(1):10634 Nishioka K, Takamoto T, Agui T, Kaneiwa M, Uraoka Y, Fuyuki T (2006) Annual output extimation of concentrator photovoltaic systems using high-efficiency InGaP/InGaAs/Ge triple-junction solar cells based on experimental solar cell’s characteristics and field-test meteorological data. Sol Energy Mater Sol Cells 90:57–67 Oh SJ, Burhan M, Ng KC, Kim Y, Chun W (2015) Development and performance analysis of a two-axis solar tracker for concentrated photovoltaics. Int J Energy Res 39(7):965–976 Oh SJ, Ng KC, Thu K, Chun W, Chua KJ (2016) Forecasting long-term electricity demand for cooling of Singapore’s buildings incorporating an innovative air-conditioning technology. Energy Build 127:183–193 Saadi A, Becherif M, Ramadan HS (2016) Hydrogen production horizon using solar energy in Biskra, Algeria. Int J Hydrog Energy 41(47):21899–21912 Shahzad MW, Burhan M, Ang L, Ng KC (2017) Energy-water-environment nexus underpinning future desalination sustainability. Desalination 413:52–64 Shahzad MW, Burhan M, Ang L, Ng KC (2018) Adsorption desalination—principles, process design, and its hybrids for future sustainable desalination. In: Emerging technologies for sustainable desalination handbook, pp 3–34 Shahzad MW, Burhan M, Ng KC (2018b) Energy storage & desalination. Int J Comput Phys Series 1(2):52–60 Shahzad MW, Burhan M, Ghaffour N, Ng KC (2018c) A multi evaporator desalination system operated with thermocline energy for future sustainability. Desalination 435:268–277 Ulleberg Ø (1998) Stand-alone power systems for the future: optimal design, operation & control of solar-hydrogen energy systems. Ph.D. Thesis, Department of Thermal Energy and Hydropower, Norwegian University of Science and Technology, Trondheim, Norway Yu X, Gen M (2010) Introduction to evolutionary algorithms. Springer Science & Business Media Chapter 13 Novel Method and Molten Salt Electrolytic Cell for Implementing a Hydrogen Fuel, Sustainable, Closed Clean Energy Cycle on a Large Scale

Alvin G. Stern

Abstract We describe an economical, novel method for implementing a hydrogen fuel clean energy cycle based on the chemical reaction between salinated (sea) or desalinated (fresh) water (H2O) and sodium (Na) metal that produces hydrogen (H2) fuel and sodium hydroxide (NaOH) byproduct. The sodium hydroxide (NaOH) is reprocessed in a solar powered electrolytic Na metal production plant that can result in excess production of chlorine (Cl2) from sodium chloride (NaCl) in sea salt mixed with NaOH, used to effect freezing point lowering of seawater reactant for hydrogen generation at reduced temperatures. The novel method and molten salt electrolytic cell enable natural separation of NaCl from NaOH, thereby limiting excess Cl2 production. The recovered NaCl can be used to produce concentrated brine solution from seawater for hydrogen generation in cold climates, or can be converted to sodium carbonate (Na2CO3) via the Solvay process for electrolytic production of Na metal without Cl2 generation.

Keywords Novel hydrogen clean energy cycle Á Novel electrolytic cell Novel electrolysis method Á Sodium metal production Á Molten salt electrolysis Solar powered electrolysis Á Safe hydrogen generation

Nomenclature cp Isobaric heat capacity (J/KÁmol) Er ; Eo Standard reduction, oxidation half reaction potential (V) Eov Standard overall reaction potential (V) ECELL Electrochemical cell potential (V) DG Change in Gibbs free energy (kJ) DGf Standard Gibbs free energy of formation (kJ/mol) DH Change in enthalpy (kJ) DHf Standard enthalpy of formation (kJ/mol)

A. G. Stern (&) AG STERN, LLC, Newton, MA 02467, USA e-mail: [email protected]

© Springer Nature Singapore Pte Ltd. 2019 277 E. Motoasca et al. (eds.), Energy Sustainability in Built and Urban Environments, Energy, Environment, and Sustainability, https://doi.org/10.1007/978-981-13-3284-5_13 278 A. G. Stern

DHfus Enthalpy of fusion (J/mol) DHtrs Enthalpy of transition (J/mol) DHvap Enthalpy of vaporization (J/mol) ICELL Electrolytic cell current (A) n Number of moles of electrons transferred (mol) P Absolute pressure (Pa) RC Electrolytic cell resistance (X) DS Change in entropy (J/K) S° Standard entropy (J/KÁmol) DSfus Entropy of fusion (J/KÁmol) DStrs Entropy of transition (J/KÁmol) DSvap Entropy of vaporization (J/KÁmol) T Absolute temperature, ITS-90 or Celsius temperature (K) or (°C) Tf Fusion temperature (K) Tb Vaporization temperature (K) VCELL Voltage applied to electrolytic cell (V) DVCELL Difference in voltage applied to electrolytic cell (V) We Electrical work (kJ) F Faraday constant [96485.3365 (C/mol)] 2 g0 Gravitational acceleration near earth’s surface [9.80665 (m/s )] P0 Standard atmospheric pressure [101325 (Pa)] T0 Celsius zero point, ITS-90 [273.15 (K)] TEu Eutectic temperature of NaCl–H2O solution [−21.2 (°C)]

13.1 Introduction

There is a need in the modern world to effect a transition to clean renewable fuels from the present large scale use of carbon based fossil fuels in transportation applications and ground based energy generation. The reasons for implementing such a transition are manifold, including first and foremost, the need to prevent irreversible environmental damage caused by the unbridled proliferation of motor vehicles equipped with hydrocarbon fueled internal combustion engines (ICEs) on a planet where the human population is inexorably increasing toward the 10 billion mark (Lutz and Samir 2010; Cohen 1995, 2003). In 2015, the carbon dioxide (CO2) mole fraction in the earth’s atmosphere for the first time attained 0.04% (400 ppm) and has been climbing steadily from a level of 315 ppm measured in 1958, when monitoring at the Mauna Loa Observatory (MLO) in Hawaii began (Dlugokencky et al. 2016). The MLO measurements were confirmed at the Cape Grim Baseline Atmospheric Pollution Station (CGBAPS) in Tasmania. Once an atmospheric mole fraction of 0.5–2% (5,000–20,000 ppm) is exceeded, human life on earth might be jeopardized (Satish et al. 2012; Guais et al. 2011). To illustrate the argument, one 13 Novel Method and Molten Salt Electrolytic Cell … 279 can recall the mass carbon dioxide poisonings that occurred on August 15, 1984 and on August 21, 1986 in Northwest Cameroon, when Lake Monoun and Lake Nyos, respectively experienced limnic eruption or lake overturn events, that suddenly released large quantities of CO2(g) dissolved in water at the bottom of the lakes. The cloud of CO2(g) gas emanating from Lake Nyos poisoned and asphyx- iated approximately 1,700 persons and 3,500 head of livestock in the surrounding villages while causing agonizing wounds to hapless survivors (Sigurdsson et al. 1987; Kling et al. 1987; Baxter et al. 1989). That such a cataclysmic event could occur naturally, had been previously unknown by scientists. The argument can be cast more starkly using an analogy based on macabre historical events. During World War 2, the three extermination camps of Operation Reinhard including Belzec, Sobibor and Treblinka, all used internal combustion engines (ICEs) fueled with liquid hydrocarbons to poison and asphyxiate countless masses of prisoners in a hermetically enclosed space called a gas chamber (Brayard 2000; Buggeln 2015). Considering that earth’s atmosphere is hermetically enclosed due to CO2(g) molecules being too heavy to escape the earth’s gravitational acceleration g0, into space, generating more CO2(g) from hydrocarbon fueled ICEs might overwhelm the capacity of vegetation to recycle it, thus creating a lethal gas chamber on the entire planet. Hydrogen (H2) which is stored in near limitless quantity in seawater is the only alternative fuel that is more abundant and environmentally cleaner with the potential of having a lower cost than nonrenewable carbon based fossil fuels, assuming that engineering challenges related to safe implementation and economical extraction of the hydrogen are overcome. The existing industrial method of generating hydrogen (H2) using steam reforming of natural gas, the latter containing mostly methane (CH4), produces significant quantities of carbon dioxide (CO2)as well as carbon monoxide (CO), even after application of the shift reaction meant to transform CO into CO2 (Xu and Froment 1989). Hydrogen is also produced commercially to a much lesser extent from water (H2O) electrolysis near room temperature using alkaline water electrolysis or proton exchange membrane (PEM) electrolysis (Marini et al. 2012; Zeng and Zhang 2010; Bessarabov et al. 2016; Carmo et al. 2013). Hydrogen generation from hydrolysis reactions between water (H2O) and chemical hydrides including sodium borohydride (NaBH4), lithium borohydride (LiBH4), lithium hydride (LiH) as well as the reactive metals lithium (Li) and aluminum (Al), continues to be studied (Muir and Yao 2011; Kojima et al. 2002, 2004a, b, c; Amendola et al. 2000; Schlesinger et al. 1953; Kong et al. 2003; Klanchar et al. 1997; Bergthorson et al. 2017). The present work describes a novel method for implementing a hydrogen fuel, sustainable, closed clean energy cycle on a large scale using a novel electrolytic cell, based on chemistry that circumvents carbon based fossil fuels and eliminates deleterious emissions of carbon monoxide (CO) and carbon dioxide (CO2). 280 A. G. Stern

13.2 Hydrogen Fuel Clean Energy Cycle Method

The novel, hydrogen fuel clean energy cycle enables a safe, reliable and cost effective means of converting solar energy into electric energy that can be supplied on demand to myriad ground based applications, by storing the sun’s radiant energy as sodium (Na) metal. The diagram of the novel, hydrogen fuel clean energy cycle is shown in Fig. 13.1. In Fig. 13.1, the complete hydrogen fuel clean energy cycle based on green chemistry is shown wherein water (H2O) and sodium (Na) metal are used to fuel a scalable, hydrogen generation apparatus that implements the controlled chemical reaction in Eq. (13.1) to produce high purity H2(g) fuel and sodium hydroxide (NaOH) byproduct (Stern 2015).

2NaðsÞ þ 2H2OðlÞ ! H2ðgÞ þ 2NaOHðsÞ ð13:1Þ

The high purity hydrogen (H2(g)) fuel produced on demand by the novel hydrogen generation apparatus can be supplied directly to internal combustion engines (ICEs) of the Otto or Diesel types or can be used to safely power fuel cell electric vehicles (FCEVs) without contaminating the sensitive platinum (Pt) catalysts present in proton exchange membrane (PEM) fuel cells or any other

Fig. 13.1 Complete hydrogen fuel clean energy cycle based on green chemistry 13 Novel Method and Molten Salt Electrolytic Cell … 281 types of catalysts in fuel cells, because the hydrogen is not derived from carbon based fossil fuels, and therefore does not contain even trace amounts of carbon monoxide or sulfur compounds (Stern 2018). The hydrogen generation apparatus uses salinated (sea) or desalinated (fresh) water (H2O) and sodium (Na) metal reactants to generate hydrogen (H2) fuel and a byproduct mixture of sodium hydroxide (NaOH) and sea salt, the latter consisting primarily of sodium chloride (NaCl) (Millero et al. 2008). The seawater reactant can be concentrated to as much as 252.18 grams of sea salt solute per kilogram of seawater solution to provide a fusion temperature TEu = −21.2 °C (251.95 K), that is equivalent to the eutectic temperature of a 23.18% by weight NaCl in NaCl–H2O solution (Bodnar 1993; Hall et al. 1988). The concentrated sea salt in seawater solution allows the hydrogen generator to operate reliably over a wide ambient temperature range from −21.2 °C (251.95 K) to 56.7 °C (329.85 K) prevailing in the conterminous 48 states of the U. S.A. In warm climates, desalinated (fresh) water (H2O) reactant can be used in the hydrogen generation apparatus to produce H2(g) fuel and sodium hydroxide (NaOH) according to Eq. (13.1). Either pure sodium hydroxide (NaOH) or a mixture of NaOH and NaCl, the latter obtained from seawater is recovered from the hydrogen generation apparatus during refueling and recycled in a self-contained solar powered electrolytic sodium (Na) metal production plant according to Eqs. (13.2) and (13.3) (Stern 2017; Castner 1891; Davy 1808; Downs 1924).

þ À 4Na þ 4OH ! 4Na þ 2H2O þ O2 ð13:2Þ

þ À 2Na þ 2Cl ! 2Na þ Cl2 ð13:3Þ

When the byproducts of the chemical reaction in Eq. (13.1) consisting primarily of NaOH and NaCl are recovered and recycled using electrolysis according to Eqs. (13.2) and (13.3), more Na metal will have been produced by the electrolysis than was originally available when the hydrogen generation apparatus was freshly fueled. Therefore, it becomes possible to passively increase the amount of sodium (Na) metal in the closed hydrogen fuel clean energy cycle. When implementing Eqs. (13.2) and (13.3) on a large scale however, toxic chlorine (Cl2) can be produced from the electrolysis of NaCl in greater quantity than can be sold to industry.

13.2.1 Selective Electrolysis to Eliminate Excess Chlorine Production

According to the cycle diagram in Fig. 13.1, the byproduct mixture of NaOH and NaCl recovered from the hydrogen generation apparatus during motor vehicle refueling is transported by truck, rail car or pipeline to self-contained solar powered electrolytic sodium (Na) metal production plants for recovery of Na metal. The electrolytic cells at the plant are charged with the aqueous mixture of NaOH(aq) and 282 A. G. Stern

NaCl(aq). The plant management then decides whether chlorine (Cl2) production is required. If the response is affirmative, then the electrolysis proceeds in a manner that results in decomposition of the entire contents of the cell including both NaOH and NaCl according to Eqs. (13.2) and (13.3), respectively to yield H2O, O2 and Cl2 at the anode and more Na metal at the cathode than had been previously used to fuel the hydrogen generation apparatus. The electrolysis of NaOH and NaCl can occur simultaneously or sequentially by using the difference in decomposition potentials between the NaOH and NaCl. The Cl2 is separated from the steam (H2O) and oxygen (O2)effluents generated at the anode of the cell to be bottled and later sold to manufacturers including paper, polymer (plastic) and chemical industries. The sodium (Na) metal produced at the cathode of the cell is hermetically packaged for reuse in generating H2(g) fuel in the hydrogen generation apparatus units. If however, Cl2 production is not required, then electrolysis proceeds at a lowered electrolytic cell voltage, in a manner that results in selective decomposition of only NaOH according to Eq. (13.2), while NaCl is not decomposed by the cell. At the end of the electrolysis, the NaCl is recovered from the electrolytic cell. According to the cycle diagram in Fig. 13.1, the plant management then decides whether the additional sodium (Na) metal from NaCl is required. If the response is negative, then the NaCl is mixed with seawater to create concentrated brine solution for use as a reactant according to Eq. (13.1) in the hydrogen generation apparatus units operating in cold climates. If the response however is affirmative, meaning that extra sodium (Na) metal from the NaCl is required to expand capacity in the hydrogen fuel clean energy cycle then the NaCl is sent for conversion to sodium carbonate (Na2CO3) in a Solvay plant that implements the net chemical reaction given in Eq. (13.4) (Steinhauser 2008).

2NaCl þ CaCO3 ! Na2CO3 þ CaCl2 ð13:4Þ

The Solvay process for the manufacture of sodium carbonate (Na2CO3) has existed since the 19th century to economically process sodium chloride (NaCl) and calcium carbonate (CaCO3) reactants, the latter also known as limestone into Na2CO3 and calcium chloride (CaCl2) products. The CaCO3 reactant is abundantly available and mined routinely while the CaCl2 byproduct of the Solvay process is not toxic to humans or deleterious to the environment and can be sold as road salt for deicing motorways in cold climates. The Na2CO3 produced by the Solvay process from NaCl recovered from the hydrogen generation apparatus can be supplied to the electrolytic cells of the self-contained solar powered electrolytic sodium (Na) metal production plants to recover the Na metal according to Eq. (13.5) (Selman and Maru 1981).

þ þ À2 ! þ þ ð : Þ 4Na 2CO3 4Na 2C 3O2 13 5

The electrolysis of Na2CO3 according to Eq. (13.5) can be achieved by operating the electrolytic cell at an elevated voltage to suppress the evolution of carbon 13 Novel Method and Molten Salt Electrolytic Cell … 283 monoxide (CO) and/or carbon dioxide (CO2) at the anode, and instead deposit pure carbon (C) and sodium (Na) metal at the cathode. The carbon deposits at the cathode of the electrolytic cell can be collected and removed. The Na2CO3 is sourced economically from mining in addition to Solvay production and therefore, constitutes a cost effective means of producing large quantities sodium (Na) metal by electrolysis while avoiding excess Cl2 accompanying production of sodium (Na) metal from electrolysis of NaCl. The hydrogen fuel clean energy cycle shown in Fig. 13.1 also supports sodium hydride (NaH) manufacture by means of a separate plant for the coproduction of elemental hydrogen (H2) in proximity to the self-contained solar powered electrolytic sodium (Na) metal production plant. The H2 required for NaH production can be obtained from clean sources using solar powered electrolysis of water (H2O) with photovoltaic (PV) panels, photocatalytic or photoelectrochemical (PEC) methods, or from biohydrogen generation using chemical, thermochemical, biological, bio- chemical and biophotolytical methods (Zhang et al. 2014). The hydrogen generation apparatus can be fueled either with Na metal or with NaH, the latter for applications where enhanced H2 generating capacity is necessary (Stern 2018). The complete hydrogen fuel clean energy cycle shown in Fig. 13.1 enables a safe, reliable and cost effective means of scaling up sodium (Na) metal production by using either Na2CO3 or NaCl for the manufacture of Na metal in the copious quantities needed for fueling hydrogen generation systems. The electrolysis of Na2CO3 provides a practical means of avoiding excess Cl2 production that might accompany the large scale electrolysis of NaCl. The hydrogen fuel, sustainable, closed clean energy cycle can be initiated using electrolysis of Na2CO3 until suf- ﬁcient Na metal exists to allow only for reprocessing of NaOH to recover the Na metal for reuse in generating H2(g) fuel, thereby avoiding Cl2 production altogether. Only if Cl2 is required by manufacturing industries, does electrolysis of NaCl have to be performed. Otherwise, the NaCl can be converted to Na2CO3 or mixed with seawater to create concentrated brine solution for use as a reactant according to Eq. (13.1) in the hydrogen generation apparatus units operating in cold climates. In Sect. 13.3, we describe the design of a novel, molten salt electrolytic cell capable of using NaOH, NaCl, Na2CO3 or a mixture of NaOH and NaCl as reactants for the production of Na metal, with selective electrolysis between NaOH and NaCl.

13.3 Molten Salt Electrolytic Cell Characteristics

The novel, molten salt electrolytic cell supports implementation of the hydrogen fuel, sustainable, closed clean energy cycle described in Fig. 13.1, and therefore must be capable of performing electrolysis on three types of molten salts individually including NaOH, NaCl, Na2CO3 or on a mixture of NaOH and NaCl, including selective electrolysis between NaOH and NaCl. The diagram of the novel electrolytic cell is shown in Fig. 13.2. 284 A. G. Stern

Fig. 13.2 Novel, molten salt electrolytic cell for enabling a hydrogen fuel, sustainable, closed clean energy cycle

The electrolytic cell shown in Fig. 13.2, incorporates novel features that enable it to function in support of the hydrogen fuel, sustainable, closed clean energy cycle. The electrolytic cell consists of a crucible section that holds the charge of fused or molten salt consisting of either pure sodium hydroxide (NaOH) or a mixture of sodium hydroxide (NaOH) and sea salt, the latter comprising primarily sodium chloride (NaCl). In addition, the crucible can hold pure sodium carbonate (Na2CO3). The crucible is heated by a radiant heater consisting of an electric heating element embedded in an insulating ceramic shell. A skirt tube made from 13 Novel Method and Molten Salt Electrolytic Cell … 285 nickel or nickel alloy capable of resisting caustic corrosion at elevated temperatures, passes through the lid of the crucible and connects the crucible to a double ended cylinder. The skirt tube rises midway into the double ended cylinder, which is located directly above the crucible lid. The skirt tube drops nearly to the floor of the crucible without contacting the bottom. A single cathode electrode rod made from nickel or nickel alloy enters from the top end of the double ended cylinder through the open skirt tube, into the crucible section of the electrolytic cell. The cathode electrode is surrounded by the skirt tube and does not extend beyond the lower end of the skirt tube into the crucible containing the molten salt(s). The cathode electrode is electrically isolated from the rest of the electrolytic cell using a ceramic fitting that prevents direct contact with the metallic parts of the cell. Multiple, concentrically arranged anode electrodes pass through the crucible lid, forming a circle around the skirt tube of the electrolytic cell. The anode electrodes can be displaced linearly along the radial direction of the circular crucible lid to vary the spacing between the anode and cathode electrodes. An increase in the distance between the electrodes results in an increase in electrical resistance RC, of the electrolytic cell while a decrease in distance yields a reduction in the electrical resistance of the cell. The anode electrodes are electrically isolated from other metal parts of the electrolytic cell using ceramic fittings. The electrolytic cell comprises 4 metal sealed bellows valves that are capable of operating reliably at elevated temperatures. The valves are used during normal operation of the electrolytic cell to introduce as well as remove cell contents including aqueous NaOH(aq), NaCl(aq), Na2CO3(aq),Na(l),H2O(g),O2(g),Cl2(g) and argon (Ar) gas. The novel electrolytic cell shown in Fig. 13.2 is designed to perform electrolysis at a temperature range between 1223.15 K (950 °C) to 1323.15 K (1050 °C) that is significantly higher than the operating temperatures of Castner or Downs type electrolytic cells where sodium (Na) metal is produced in a liquid state during electrolysis (Castner 1891; Downs 1924). There exist two principal reasons for operating the novel electrolytic cell at such elevated temperatures. First, the high temperature operation enables the fused salt(s) in the crucible to become completely anhydrous by evaporating any residual water (H2O) content from the salt(s) including the water produced at the cell anode. Second, the high temperature of the electrolytic cell is well above the boiling point temperature of sodium (Na) metal given as Tb = 1154.5 ± 4.7 K (Makansi et al. 1955). As a result, the sodium (Na) metal is produced in a vapor state at the cathode and does not dissolve into the fused salt(s), thereby providing a direct means of separating and purifying the Na metal where otherwise, the Na produced at the cathode might dissolve into the fused salt(s) and react with steam (H2O(g)) and oxygen (O2(g)) produced at the anode to lower the overall efficiency and production yield of Na metal by the electrolytic cell (Von Hevesy 1909; Lorenz and Winzer 1929). The novel electrolytic cell will implement electrolysis individually on fused NaOH(l), NaCl(l),Na2CO3(l) or on a mixture of NaOH(l) and NaCl(l), with selective electrolysis between NaOH(l) and NaCl(l) according to Eqs. (13.6)–(13.8). 286 A. G. Stern

þ À 4Na þ 4OH ! 4NaðgÞ þ 2H2OðgÞ þ O2ðgÞ ð13:6Þ

þ À 2Na þ 2Cl ! 2NaðgÞ þ Cl2ðgÞ ð13:7Þ

þ þ À2 ! þ þ ð : Þ 4Na 2CO3 4NaðgÞ 2CðsÞ 3O2ðgÞ 13 8

In Eqs. (13.6)–(13.8), the sodium (Na) metal produced at the cell cathode in a vapor state Na(g), rises up through the riser tube over the fused salts, subsequently expanding into the double ended cylinder located directly above the crucible lid, whereupon it condenses to liquid Na(l) to be collected. The novel electrolytic cell shown in Fig. 13.2 is meant to be operated as an integral component of scalable, self-contained solar powered electrolytic sodium (Na) metal production plants with up to 25 such cells electrically connected together in series (Stern 2017).

13.3.1 Electrolysis of Pure NaOH

The method of operation for the novel electrolytic cell depends on the chemical contents of the crucible. If the hydrogen generation apparatus indicated in Fig. 13.1 was originally fueled with desalinated (fresh) water (H2O) and sodium (Na) metal reactants to generate hydrogen (H2) fuel and sodium hydroxide (NaOH) byproduct, then the electrolytic cell will perform electrolysis on pure NaOH. The electrolytic cell operation commences before dawn by first opening valve 1 and filling the crucible with aqueous NaOH(aq), followed by the closing of valve 1. At sunrise, valves 2 and 3 are opened and the electric current generated by the photovoltaic (PV) device panel array is directed to the radiant electric heater shown in Fig. 13.2 to fuse or melt the NaOH and elevate the temperature of the NaOH(l) to range between 1223.15 K (950 °C) to 1323.15 K (1050 °C). Steam (H2O(g)) released from the aqueous NaOH(aq) can be vented out from the cell through open valves 2 and 3. After venting steam, valve 4 is opened to allow argon (Ar) gas introduced through valve 4 to purge the interior of the riser tube and double ended cylinder of any air and water vapor that could react chemically with sodium (Na(g)) metal vapor. The Ar gas flushes out the air and moisture through valve 3 that has remained open. Once the current from the photovoltaic (PV) device panel array reaches a minimum threshold sufficient for electrolysis, valve 4 is closed and valve 3 continues remaining open in preparation to commence electrolysis. During electrolysis according to Eq. (13.6), steam (H2O(g)) and oxygen (O2(g)) are produced at the anode and exit the electrolytic cell through open valve 2. The sodium (Na) metal is produced at the cathode in a vapor state Na(g), rises up through the riser tube over the fused NaOH(l), subsequently expanding into the double ended cylinder located directly above the crucible lid, whereupon it condenses to liquid Na(l) to be transferred from the double ended cylinder through valve 3 for 13 Novel Method and Molten Salt Electrolytic Cell … 287

Table 13.1 Fusion and vaporization temperatures at P0 = 101325 Pa

Fusion temperature Tf (K) Vaporization temperature Tb (K) Sodium Hydroxide 594 ± 2 (Gurvich et al. 1996) 1661.15 (Wartenberg and (NaOH) Albrecht 1921) Sodium (Na) 370.95 ± 0.05 (Alcock et al. 1994) 1154.5 ± 4.7 (Makansi et al. 1955) aPure (VSMOW) 273.152518 ± 0.000002 (Feistel and 373.124 (Wagner and Prub Water (H2O) Wagner 2005, 2006) 2002)

Oxygen (O2) 54.37 (Stewart et al. 1991) 90.19 (Stewart et al. 1991) aAir free water as opposed to air-saturated water, the latter having a freezing point T = 273.1501 K (Doherty and Kester 1974)

Table 13.2 Enthalpies of fusion and vaporization at P0 = 101325 Pa a b Enthalpy of fusion DHfus, DHtrs (J/mol) Enthalpy of vaporization DHvap (J/mol) Sodium Hydroxide 6360 ± 400 (Gurvich et al. 1996), (NaOH) 0 (Gurvich et al. 1996) 6360 ± 300 (Gurvich et al. 1996) Sodium (Na) 2600 ± 25 (Alcock et al. 1994) 89240.15 (Fink and Leibowitz 1995) Pure (VSMOW) 6006.83 (Feistel and Wagner 2006) 40650.95 (Wagner and Prub Water (H2O) 2002) a Enthalpy of transition DHtrs from a ! b (514 K) and b ! c (568 K) bEnthalpy of vaporization at the vaporization temperature provided in Table 13.1 packaging and reuse in the hydrogen generation apparatus units. Table 13.1 provides the fusion and vaporization temperatures for the reactants and products in Eq. (13.6). Table 13.2 provides the enthalpies of fusion, transition and vaporization for the reactants and products in Eq. (13.6). Table 13.3 provides the entropies of fusion, transition and vaporization for the reactants and products in Eq. (13.6). The voltage required for electrolysis of pure NaOH(l) according to Eq. (13.6), can be estimated from the standard reduction potentials of the oxidation and reduction half reactions that occur at the anode and cathode, respectively of the electrolysis cell (Weast 1976).

ðÞ: þ þ À ! ¼À : ð : Þ Cathode reduction NaðaqÞ e NaðsÞ Er 2 71 V 13 9

ðÞ: À ! þ þ À ¼À : ð : Þ Anode oxidation 4OHðaqÞ O2ðgÞ 2H2O 4e Eo 0 40 V 13 10 288 A. G. Stern

Table 13.3 Entropies of fusion and vaporization at P0 = 101325 Pa a a Entropy of fusion DSfus, Entropy of vaporization DSvap (J/ DStrs (J/KÁmol) KÁmol) Sodium Hydroxide 10.70 (Gurvich et al. 1996), (NaOH) 11.197 (Gurvich et al. 1996) Sodium (Na) 7.00 (Alcock et al. 1994) 77.298 (Makansi et al. 1955; Fink and Leibowitz 1995) Pure (VSMOW) 21.99 (Feistel and Wagner 108.947 (Wagner and Prub 2002) Water (H2O) 2005, 2006) aEntropy of fusion, transition, and vaporization calculated using DS = DH / T, with corresponding enthalpies and temperatures

From Eqs. (13.9) and (13.10), the minimum potential of Eov ¼À3:11 V is needed to electrolyze NaOH. The standard reduction potentials of the oxidation and reduction half reactions however, are applicable only for a 1 M concentration aqueous solution of NaOH(aq) at temperature T = 298.15 K and atmospheric pressure P0 = 101325 Pa. Therefore, the minimum cell potential Eov ¼À3:11V cannot be considered valid for the electrolytic decomposition of fused or molten NaOH(l) at the cell operating temperature ranging between 1223.15 K (950 °C) to 1323.15 K (1050 °C) described by Eq. (13.6). To calculate an accurate estimate of the decomposition potential for fused NaOH(l) in the novel electrolytic cell according to Eq. (13.6) at the higher temperatures, it is necessary to use Eq. (13.11).

DG ¼Àn Á F Á ECELL ð13:11Þ

In Eq. (13.11), DG represents the change in the Gibbs free energy of the chemical reaction and corresponds to the maximum useful electrical work that can be obtained from the reaction given as We = −DG. We utilize the convention that if the Gibbs free energy DG of a reaction is negative, then useful electrical work We, produced by the reaction will be positive. In Eq. (13.11), n represents the number of moles of electrons transferred through the circuit and the Faraday constant F, represents the electric charge of 1 mol of electrons in Coulombs. The negative sign in Eq. (13.11) serves to make the electrochemical cell potential ECELL, positive if DG of the reaction is negative, meaning that it functions as a battery to produce useful electrical work We.IfDG of the reaction is positive, then the electrochemical cell potential ECELL, will be negative, thus functioning as an electrolytic cell that requires external energy input. The change in the Gibbs free energy of a chemical reaction at constant temperature is given by Eq. (13.12) (Gibbs 1878).

DG ¼ DH À TDS ð13:12Þ 13 Novel Method and Molten Salt Electrolytic Cell … 289

Table 13.4 Gibbs free energy, enthalpy of formation, and entropy at T = 298.15 K and P0 = 101325 Pa ° DGf (kJ/mol) DHf (kJ/mol) S (J/KÁmol) Sodium Hydroxide −419.17 (Weast −469.61 (Weast 49.79 (Weast 1976) (NaOH(aq)) 1976) 1976) Sodium Hydroxide −379.651 (Gurvich −425.8 ± 0.2 64.43 ± 0.8 (NaOH(s)) et al. 1996) (Gurvich et al. 1996) (Gurvich et al. 1996)

Sodium (Na(s)) 0 0 51.1 ± 0.3 (Alcock et al. 1994) Pure (VSMOW) −237.19 (Weast −285.85 (Weast 69.96 (Weast 1976) Water (H2O(l)) 1976) 1976)

Oxygen (O2(g)) 0 0 205.03 (Weast 1976)

In Eq. (13.12), DG represents the change in the Gibbs free energy, DH represents the enthalpy change, and DS represents the entropy change of the reaction. Table 13.4 lists the standard Gibbs free energies of formation DGf , standard ° enthalpies of formation DHf , and standard entropies S of the compounds in Eqs. (13.9) and (13.10) at temperature T = 298.15 K and standard atmospheric pressure of P0 = 101325 Pa. Using the tabulated standard free energies of formation DGf , in Table 13.4,it becomes possible to calculate the change in the Gibbs free energy for the electrolytic decomposition of aqueous sodium hydroxide (NaOH(aq)) according to Eqs. (13.9) and (13.10), as shown below in Eq. (13.13). ÀÁÀÁ ÀÁ ÀÁ D ¼ Á D þ Á D þ D Àð G 4 GÀÁf NaðsÞ 2 Gf H2OðlÞ Gf O2ðgÞ 4 Á DGf NaOHðaqÞ Þ ¼ 1202:3kJ ð13:13Þ

The result of DG = 1202.3 kJ from Eq. (13.13) using aqueous NaOH(aq) and water in a liquid state H2O(l) is valid at temperature T = 298.15 K and pressure P0 = 101325 Pa. The large positive value for DG shows that under these conditions, the chemical decomposition of the NaOH(aq) will not be spontaneous and instead, require an external input of energy. Using Eq. (13.11) with the result from Eq. (13.13), it is possible to calculate the potential required for the electrolytic cell to decompose the NaOH(aq), as shown in Eq. (13.14).

ECELL ¼ DG=ðÀn Á FÞ¼1; 202; 300 J =ðÀ4 Á 96; 485:3365 C=molÞ¼À3:11 V ð13:14Þ 290 A. G. Stern

The potential of the electrolytic cell ECELL = −3.11 V, needed to decompose NaOH(aq) matches exactly the result obtained from the standard reduction potentials of the oxidation and reduction half reactions that occur at the anode and cathode, respectively of the electrolysis cell described in Eqs. (13.9) and (13.10), as might be expected.

13.3.1.1 Electrolysis of Pure NaOH as a Function of Temperature

To calculate the potential of the electrolytic cell needed to decompose fused or molten NaOH(l) at temperatures ranging between 1223.15 K (950 °C) to 1323.15 K (1050 °C) described by Eq. (13.6), it is necessary to calculate the entropy values of the reactants and products listed in Table 13.1, as well as the enthalpies of formation of the NaOH(l) and H2O(g) as a function of temperature. The Eq. (13.12) can then be used to calculate the resulting change in the Gibbs free energy for the electrochemical decomposition reaction in Eq. (13.6) as a function of temperature. The calculation of the change in entropy as a function of temperature for the reactants and products in Eq. (13.6) is performed according to Eq. (13.15).

TZ2 c DS ¼ p dT ð13:15Þ T T1

The calculation of the change in the enthalpies of formation as a function of temperature for NaOH(l) and H2O(g), is performed according to Eq. (13.16).

TZ2 DH ¼ DcpdT ð13:16Þ T1

In Eq. (13.16), DH represents the difference in the enthalpy of formation between the values at T1 and T2, while Dcp represents the difference in the isobaric heat capacities between products and reactants. Equation (13.16) is used to evaluate the chemical reactions provided in Eqs. (13.17) and (13.18).

NaðgÞ þ ðÞ1=2 O2ðgÞ þ ðÞ1=2 H2ðgÞ ! NaOHðlÞ ð13:17Þ

H2ðgÞ þ ðÞ1=2 O2ðgÞ ! H2OðgÞ ð13:18Þ

Table 13.5 Heat capacities at constant pressure with P0 = 101325 Pa

Temperature cp (J/KÁmol) (K) Sodium 298.15 59.53 ± 0.3 (Gurvich et al. 1996) Hydroxide a-NaOH: −797.602 + 5146.32Â10−3 Á T + 87.154Â105 Á T−2 − (NaOH) 298.15–514 11367.7Â10−6 Á T2 + 8875.35Â10−9 Á T3 (Gurvich et al. 1996) b-NaOH: 80 (Gurvich et al. 1996) 514–568 c-NaOH: 86 (Gurvich et al. 1996) 568–594 594–1000 89.613 − 5.913Â10−3 Á T (Gurvich et al. 1996) 1000–2000 83.7 (Gurvich et al. 1996) Sodium (Na) 298.15 28.16 ± 0.1 (Alcock et al. 1994) 298.15– 53.941 − 194.429Â10−3 Á T + 361.905Â10−6 Á T2 370.95 (Alcock et al. 1994) 370.95–1600 37.466 − 19.148Â10−3 Á T + 10.628Â10−6 Á T2 (Alcock et al. 1994)

The isobaric heat capacity cp, as a function of temperature is provided in Table 13.6 for hydrogen (H2), oxygen (O2) and pure (VSMOW) water (H2O). Using Eq. (13.15) with the isobaric heat capacities given in Tables 13.5 and 13.6, as well as the entropies of fusion, transition and vaporization given in Table 13.3, it becomes possible to calculate the entropy values of the reactants and products in Eq. (13.6) as a function of the temperature. In Fig. 13.3, the entropies as a function of temperature are calculated for sodium hydroxide (NaOH), sodium (Na), pure (VSMOW) water (H2O) and oxygen (O2). In Fig. 13.4, the enthalpies of formation as a function of temperature are calculated for sodium hydroxide (NaOH) and pure (VSMOW) water (H2O) taking into consideration the enthalpies of fusion, transition and vaporization given in Table 13.2. In Fig. 13.5, using the calculated entropies and enthalpies of formation as a function of temperature shown in Figs. 13.3 and 13.4, respectively, it becomes possible to calculate the change in the Gibbs free energy of the electrochemical decomposition reaction in Eq. (13.6) as a function of the temperature, according to Eq. (13.12), and subsequently the electrochemical potential ECELL, using Eq. (13.11). The calculation in Fig. 13.5 shows that applying a potential VCELL = 1.78 V will be sufﬁcient to electrolyze NaOH(l) at temperatures ranging between 1223.15 K (950 °C) to 1323.15 K (1050 °C). 292 A. G. Stern

Table 13.6 Heat capacities at constant pressure with P0 = 101325 Pa

Temperature (K) cp (J/KÁmol)

Hydrogen (H2) 300 28.85 (Hilsenrath 1955) 400 29.19 (Hilsenrath 1955) 500 29.26 (Hilsenrath 1955) 600 29.33 (Hilsenrath 1955) 700 29.44 (McBride et al. 1963) 800 29.63 (McBride et al. 1963) 900 29.88 (McBride et al. 1963) 1000 30.21 (McBride et al. 1963) 1100 30.58 (McBride et al. 1963) 1200 30.99 (McBride et al. 1963) 1300 31.43 (McBride et al. 1963) 1400 31.87 (McBride et al. 1963)

Oxygen (O2) 300 29.44 (Stewart et al. 1991) 400 30.13 (Hilsenrath 1955) 500 31.11 (Hilsenrath 1955) 600 32.10 (Hilsenrath 1955) 700 32.99 (Hilsenrath 1955) 800 33.74 (Hilsenrath 1955) 900 34.37 (Hilsenrath 1955) 1000 34.88 (Hilsenrath 1955) 1100 35.31 (Hilsenrath 1955) 1200 35.68 (Hilsenrath 1955) 1300 36.00 (Hilsenrath 1955) 1400 36.30 (Hilsenrath 1955)

Pure (VSMOW) Water (H2O) 300 75.31 (Wagner and Prub 2002) 350 75.57 (Wagner and Prub 2002) a373.124 75.95 (Wagner and Prub 2002) b373.124 37.47 (Wagner and Prub 2002) 400 36.20 (Wagner and Prub 2002) 500 35.70 (Wagner and Prub 2002) 600 36.52 (Wagner and Prub 2002) 700 37.59 (Wagner and Prub 2002) 800 38.78 (Wagner and Prub 2002) 900 40.02 (Wagner and Prub 2002) 1000 41.29 (Wagner and Prub 2002) 1100 42.55 (Wagner and Prub 2002) 1200 43.78 (Wagner and Prub 2002) 1273 44.64 (Wagner and Prub 2002) aVaporization temperature, (liquid state) bVaporization temperature, (vapor state) 13 Novel Method and Molten Salt Electrolytic Cell … 293

Fig. 13.3 Entropy as a function of temperature for sodium hydroxide (NaOH) (thin solid), sodium (Na) (thick solid), pure (VSMOW) water (H2O) (thin dash) and oxygen (O2) (thick dash)

Fig. 13.4 Enthalpy of formation as a function of temperature for sodium hydroxide (NaOH) (thin solid) and pure (VSMOW) water (H2O) (thin dash)

Fig. 13.5 Electrochemical potential ECELL of the electrolytic cell as a function of temperature for the reaction in Eq. (13.6) 294 A. G. Stern 13.3.2 Electrolysis of a Mixture of NaOH and NaCl

If the hydrogen generation apparatus indicated in Fig. 13.1 was originally fueled with salinated (sea) water (H2O) and sodium (Na) metal reactants to generate hydrogen (H2) fuel, then the electrolytic cell will perform electrolysis on a mixture of NaOH and sea salt, the latter consisting primarily of sodium chloride (NaCl). If the seawater reactant was concentrated up and 252.18 grams of sea salt solute per kilogram of seawater solution, then the electrolytic cell will contain a mixture of 13.19% by weight sea salt and 86.81% by weight NaOH. The electrolytic cell operation commences before dawn by first opening valve 1 and filling the crucible with aqueous NaOH(aq) and NaCl(aq), followed by the closing of valve 1. At sunrise, valves 2 and 3 are opened and the electric current generated by the photovoltaic (PV) device panel array is directed to the radiant electric heater shown in Fig. 13.2 to fuse or melt the NaOH and NaCl mixture and elevate the temperature of the NaOH(l) and NaCl(l) to range between 1223.15 K (950 °C) to 1323.15 K (1050 °C). Steam (H2O(g)) released from the aqueous NaOH(aq) and NaCl(aq) can be vented out from the cell through open valves 2 and 3. After venting steam, valve 4 is opened to allow argon (Ar) gas introduced through valve 4 to purge the interior of the riser tube and double ended cylinder of any air and water vapor that could react chemically with sodium (Na(g)) metal vapor. The Ar gas flushes out the air and moisture through valve 3 that has remained open. Once the current from the photovoltaic (PV) device panel array reaches a minimum threshold sufficient for electrolysis, valve 4 is closed and valve 3 continues remaining open in preparation to commence electrolysis. At this point, the plant management must consult the hydrogen fuel, sustainable, closed clean energy cycle diagram shown in Fig. 13.1 and determine if chlorine (Cl2) production is required. If the answer is affirmative, then there are two possible modes of operating the electrolytic cell to yield Cl2(g) from NaCl, either using simultaneous electrolysis of NaOH and NaCl, or using selective electrolysis between NaOH and NaCl. The voltage required for electrolysis of pure NaOH has been calculated in the previous section (Sect. 13.3.1.1). During simultaneous electrolysis of NaOH and NaCl according to Eqs. (13.6) and (13.7), steam (H2O(g)), oxygen (O2(g)) and chlorine (Cl2(g)) are produced at the anode and exit the electrolytic cell through open valve 2. The sodium (Na) metal is produced at the cathode in a vapor state Na(g), rises up through the riser tube over the mixture of fused NaOH(l) and NaCl(l), subsequently expanding into the double ended cylinder located directly above the crucible lid, whereupon it condenses to liquid Na(l) to be transferred from the double ended cylinder through valve 3 for packaging and reuse in the hydrogen generation apparatus units. Table 13.7 provides the fusion and vaporization temperatures for the reactants and products in Eq. (13.7). Table 13.8 provides the enthalpies of fusion and vaporization for the reactants and products in Eq. (13.7). Table 13.9 provides the entropies of fusion and vaporization for the reactants and products in Eq. (13.7). 13 Novel Method and Molten Salt Electrolytic Cell … 295

Table 13.7 Fusion and vaporization temperatures at P0 = 101325 Pa

Fusion temperature Tf (K) Vaporization temperature Tb (K) Sodium 1073.75 (Akella et al. 1969) 1712.15 (Wartenberg and Chloride (NaCl) Albrecht 1921) Sodium (Na) 370.95 ± 0.05 (Alcock et al. 1994) 1154.5 ± 4.7 (Makansi et al. 1955) a Chlorine (Cl2) 172.17 ± 0.05 (Arblaster 2013; Giauque 239.18 (Arblaster 2013; and Powell 1939) Ambrose 1979) aFusion temperature represents the triple point at P = 1392 Pa; Vaporization temperature at P0 = 101325 Pa

Table 13.8 Enthalpies of fusion and vaporization at P0 = 101325 Pa a Enthalpy of fusion DHfus Enthalpy of vaporization DHvap (J/mol) (J/mol) Sodium Chloride 27990.96 (Dworkin and Bredig (NaCl) 1960) Sodium (Na) 2600 ± 25 (Alcock et al. 1994) 89240.15 (Fink and Leibowitz 1995) aEnthalpy of vaporization at the vaporization temperature provided in Table 13.7

Table 13.9 Entropies of fusion and vaporization at P0 = 101325 Pa a a Entropy of fusion DSfus Entropy of vaporization DSvap (J/KÁmol) (J/KÁmol) Sodium Chloride 26.07 (Akella et al. 1969; (NaCl) Dworkin and Bredig 1960) Sodium (Na) 7.00 (Alcock et al. 1994) 77.298 (Makansi et al. 1955; Fink and Leibowitz 1995) a Entropy of fusion and vaporization calculated using the formulas DSfus = DHfus / Tf and DSvap = DHvap / Tb, respectively

The voltage required for electrolysis of pure NaCl(l) according to Eq. (13.7), can be estimated from the standard reduction potentials of the oxidation and reduction half reactions that occur at the anode and cathode, respectively of the electrolysis cell (Weast 1976).

ðÞ: þ þ À ! ¼À : ð : Þ Cathode reduction NaðaqÞ e NaðsÞ Er 2 71 V 13 19 ðÞ: À ! þ À ¼À : ð : Þ Anode oxidation 2ClðaqÞ Cl2ðgÞ 2e Eo 1 36 V 13 20

From Eqs. (13.19) and (13.20), the minimum potential of Eov ¼À4:07 V is needed to electrolyze NaCl. The standard reduction potentials of the oxidation and 296 A. G. Stern

Table 13.10 Gibbs free energy, enthalpy of formation and entropy at T = 298.15 K and P0 = 101325 Pa ° DGf (kJ/mol) DHf (kJ/mol) S (J/KÁmol) Sodium Chloride −393.04 (Weast, −407.11 (Weast 115.48 (Weast 1976) (NaCl(aq)) 1976) 1976) Sodium Chloride −384.28 (Archer −411.27 (Archer 72.27 (Archer 1992) (NaCl(s)) 1992) 1992)

Sodium (Na(s)) 0 0 51.1 ± 0.3 (Alcock et al. 1994)

Chlorine (Cl2(g)) 0 0 222.97 (Weast 1976) reduction half reactions however, are applicable only for a 1 M concentration aqueous solution of NaCl(aq) at temperature T = 298.15 K and atmospheric pressure P0 = 101325 Pa. Therefore, the minimum cell potential Eov = −4.07 V cannot be considered valid for the electrolytic decomposition of fused or molten NaCl(l) at the cell operating temperature ranging between 1223.15 K (950 °C) to 1323.15 K (1050 °C) described by Eq. (13.7). A similar procedure as described in Sect. 13.3.1 can be used to calculate an accurate estimate of the decomposition potential ECELL, for fused NaCl(l) in the novel electrolytic cell according to Eq. (13.7) at the higher temperatures using Eqs. (13.11) and (13.12). Table 13.10 lists the standard Gibbs free energies of formation DGf , standard enthalpies of formation DHf , and standard entropies S° of the compounds in Eqs. (13.19) and (13.20) at temperature T = 298.15 K and standard atmospheric pressure of P0 = 101325 Pa. Using the tabulated standard free energies of formation DGf , in Table 13.10,it becomes possible to calculate the change in the Gibbs free energy for the electrolytic decomposition of aqueous sodium chloride (NaCl(aq)) according to Eqs. (13.19) and (13.20), as shown in Eq. (13.21). ÀÁÀÁ ÀÁÀÁÀÁ DG ¼ 2 Á DGf NaðsÞ þ DGf Cl2ðgÞ À 2 Á DGf NaClðaqÞ ¼ 786:08 kJ ð13:21Þ

The result of DG = 786.08 kJ from Eq. (13.21) using aqueous NaCl(aq) is valid at temperature T = 298.15 K and pressure P0 = 101325 Pa. The large positive value for DG shows that under these conditions, the chemical decomposition of the NaCl(aq) will not be spontaneous and instead, require an external input of energy. Using Eq. (13.11) with the result from Eq. (13.21), it is possible to calculate the potential required for the electrolytic cell to decompose the NaCl(aq), as shown in Eq. (13.22).

ECELL ¼ DG=ðÀn Á FÞ¼786080 J =ðÀ2 Á 96485:3365 C=molÞ¼À4:07 V ð13:22Þ

The potential of the electrolytic cell ECELL = −4.07 V, needed to decompose NaCl(aq) matches exactly the result obtained from the standard reduction potentials 13 Novel Method and Molten Salt Electrolytic Cell … 297 of the oxidation and reduction half reactions that occur at the anode and cathode, respectively of the electrolysis cell described in Eqs. (13.19) and (13.20), as might be expected.

13.3.2.1 Electrolysis of a Mixture of NaOH and NaCl as a Function of Temperature

To calculate the potential of the electrolytic cell needed to decompose fused or molten NaCl(l) at temperatures ranging between 1223.15 K (950 °C) and 1323.15 K (1050 °C) described by Eq. (13.7), it is necessary to calculate the entropy values of the reactants and products listed in Table 13.7, as well as the enthalpy of formation of the NaCl(l) as a function of temperature. Equation (13.12) can then be used to calculate the resulting change in the Gibbs free energy for the electrochemical decomposition reaction in Eq. (13.7) as a function of temperature. The calculation of the change in entropy as a function of temperature for the reactants and products in Eq. (13.7), is performed according to Eq. (13.15). The change in the enthalpy of formation for NaCl(l), is calculated according to Eq. (13.16) by evaluating the chemical reaction provided in Eq. (13.23).

NaðgÞ þ ðÞ1=2 Cl2ðgÞ ! NaClðlÞ ð13:23Þ

Calculating the entropy and the enthalpy of formation at different temperatures ranging between 1223.15 K (950 °C) to 1323.15 K (1050 °C) requires knowledge of the isobaric heat capacity cp, as a function of temperature which is provided in Table 13.11 for sodium chloride (NaCl) and chlorine (Cl2). Using Eq. (13.15) in conjunction with the isobaric heat capacities given in Tables 13.5 and 13.11, as well as the entropies of fusion and vaporization given in Table 13.9, it becomes possible to calculate the entropy values of the reactants and products in Eq. (13.7) as a function of the temperature. In Fig. 13.6, the entropies as a function of temperature are calculated for sodium chloride (NaCl), sodium (Na), and chlorine (Cl2). In Fig. 13.7, the enthalpy of formation as a function of temperature is calculated for sodium chloride (NaCl) taking into consideration the enthalpies of fusion and vaporization given Table 13.8. In Fig. 13.8, using the calculated entropies and enthalpies of formation as a function of temperature shown in Figs. 13.6 and 13.7, respectively, it becomes possible to calculate the change in the Gibbs free energy of the electrochemical decomposition reaction in Eq. (13.7) as a function of the temperature, according to Eq. (13.12), and subsequently the electrochemical potential ECELL, using Eq. (13.11). The calculation in Fig. 13.8 shows that applying a potential VCELL = 3.19 V will be sufﬁcient to electrolyze NaCl(l) at temperatures ranging between 1223.15 K (950 °C) to 1323.15 K (1050 °C). To electrolyze the mixture of fused NaOH(l) and NaCl(l) simultaneously, the voltage of the electrolytic cell is set to VCELL 3.19 V to yield H2O(g),O2(g) and 298 A. G. Stern

Table 13.11 Heat capacities at constant pressure with P0 = 101325 Pa

Temperature (K) cp (J/KÁmol) Sodium Chloride (NaCl) 298.15 50.16 (Archer 1992) 304.40 50.33 (Leadbetter and Settatree 1969) 352.80 51.34 (Leadbetter and Settatree 1969) 392.50 52.26 (Leadbetter and Settatree 1969) 464.20 53.39 (Leadbetter and Settatree 1969) 493.30 53.85 (Leadbetter and Settatree 1969) 548.60 54.77 (Leadbetter and Settatree 1969) 602.30 55.40 (Leadbetter and Settatree 1969) 649.50 55.90 (Leadbetter and Settatree 1969) 700.10 57.36 (Leadbetter and Settatree 1969) 772.50 58.99 (Leadbetter and Settatree 1969) a1073.75–1712.15 71 ± 6 (Redkin et al. 2015)

Chlorine (Cl2) 298.15 33.94 (McBride et al. 1963) 300 33.97 (McBride et al. 1963) 400 35.30 (McBride et al. 1963) 500 36.09 (McBride et al. 1963) 600 36.58 (McBride et al. 1963) 700 36.92 (McBride et al. 1963) 800 37.16 (McBride et al. 1963) 900 37.35 (McBride et al. 1963) 1000 37.49 (McBride et al. 1963) 1100 37.62 (McBride et al. 1963) 1200 37.73 (McBride et al. 1963) 1300 37.82 (McBride et al. 1963) 1400 37.91 (McBride et al. 1963) a Isobaric heat capacity for molten NaCl(l) is reported to be temperature independent (Redkin et al. 2015)

Fig. 13.6 Entropy as a function of temperature for sodium chloride (NaCl) (thin solid), sodium (Na) (thick solid) and chlorine (Cl2) (thin dash) 13 Novel Method and Molten Salt Electrolytic Cell … 299

Fig. 13.7 Enthalpy of formation as a function of temperature for sodium chloride (NaCl)

Fig. 13.8 Electrochemical potential ECELL, of the electrolytic cell as a function of temperature for the reaction in Eq. (13.7)

Cl2(g) at the anode and Na(g) at the cathode. It is preferable however, to perform selective electrolysis between fused NaOH(l) and NaCl(l), favored at high temperatures above the boiling point of sodium (Na) metal where the difference in the decomposition potentials between the fused NaCl(l) and NaOH(l) is given as DVCELL = 3.19 V − 1.78 V = 1.41 V at the boiling point temperature of sodium (Na) metal Tb = 1154.5 ± 4.7 K (Makansi et al. 1955). The unique mode of selective electrolysis operation of the novel electrolytic cell is achieved by using elevated cell temperatures ranging between 1223.15 K (950 °C) to 1323.15 K (1050 °C) above the boiling point of Na metal, ensuring that overvoltages at the anode and cathode terminals are reduced to almost zero. Thus, setting the voltage of the electrolytic cell to VCELL 1.78 V initially, ensures that only the fused NaOH(l) will be electrolyzed according to Eq. (13.6). Once the NaOH(l) has been consumed 300 A. G. Stern in the electrolytic cell and no further evolution of H2O(g) and O2(g) occurs at the anode, then the voltage of the cell can be increased to VCELL 3.19 V to electrolyze the remaining fused NaCl(l) according to Eq. (13.7). Selective electrolysis offers the principal advantage of naturally separating the electrolysis products at the cell anode producing pure steam (H2O(g)) mixed with oxygen (O2(g)), followed by production of pure Cl2(g). The plant management consults the diagram shown in Fig. 13.1, and if it determines that Cl2 production is not required, then selective electrolysis of only NaOH can occur according to Eq. (13.6) in the electrolytic cell. The NaCl remaining in the cell after selective electrolysis of the NaOH is completed, can be ﬂushed out from the crucible by dissolving in seawater which is pumped into the crucible through valve 1. The concentrated seawater containing the dissolved aqueous NaCl(aq) can be pumped out from the cell through valve 4 and stored in a tank for later use. The plant management once again consults the diagram shown in Fig. 13.1 and determines if the sodium (Na) metal contained in the aqueous NaCl(aq) recovered from the electrolytic cell as a concentrated solute in seawater is required to increase the quantity of the Na metal in the hydrogen fuel, sustainable, closed clean energy cycle. If the response is negative, meaning that extra Na metal is not required, then the concentrated sea salt in seawater solution can be used as a reactant for the hydrogen generation apparatus units operating in cold climates as indicated in Fig. 13.1. If however, the Na metal from the NaCl is required, then the concentrated seawater solution can be transported to a Solvay plant as shown in Fig. 13.1 for conversion to sodium carbonate (Na2CO3). The Na2CO3 can subsequently be returned to the self-contained solar powered electrolytic sodium (Na) metal production plant in the form of an aqueous Na2CO3(aq) solution for electrolysis, to recover Na metal without Cl2 production according to Eq. (13.8).

13.3.3 Electrolysis of Na2CO3

The electrolysis of sodium carbonate (Na2CO3) is meant occur when the hydrogen fuel, sustainable, closed clean energy cycle is first initiated, until sufficient Na metal exists to allow only for reprocessing of NaOH to recover the Na metal for reuse in generating H2(g) fuel. The Na2CO3 can be sourced from mining or from conversion of the NaCl collected from seawater, in a Solvay plant as indicated in Fig. 13.1. The electrolysis of Na2CO3 according to Eq. (13.8), allows Na metal to be produced on a large scale for the hydrogen fuel, sustainable, closed clean energy cycle while avoiding excess Cl2 production that might accompany the large scale electrolysis of sodium chloride (NaCl) collected from seawater. The electrolytic cell operation commences before dawn by first opening valve 1 and filling the crucible with aqueous Na2CO3(aq), followed by the closing of valve 1. At sunrise, valves 2 and 3 are opened and the electric current generated by the photovoltaic (PV) device panel array is directed to the radiant electric heater shown in Fig. 13.2 to fuse or melt the 13 Novel Method and Molten Salt Electrolytic Cell … 301

Na2CO3 and elevate the temperature of the Na2CO3(l) to range between 1223.15 K (950 °C) to 1323.15 K (1050 °C). Steam (H2O(g)) released from the aqueous Na2CO3(aq) can be vented out from the cell through open valves 2 and 3. After venting steam, valve 4 is opened to allow argon (Ar) gas introduced through valve 4 to purge the interior of the riser tube and double ended cylinder of any air and water vapor that could react chemically with sodium (Na(g)) metal vapor. The Ar gas ﬂushes out the air and moisture through valve 3 that has remained open. Once the current from the photovoltaic (PV) device panel array reaches a minimum threshold sufﬁcient for electrolysis, valve 4 is closed and valve 3 continues remaining open in preparation to commence electrolysis. During electrolysis of Na2CO3 according to Eq. (13.8), oxygen (O2(g)) is produced at the anode and exits the electrolytic cell through open valve 2. The sodium (Na) metal is produced at the cathode in a vapor state Na(g), rises up through the riser tube over the fused Na2CO3(l), subsequently expanding into the double ended cylinder located directly above the crucible lid, whereupon it condenses to liquid Na(l) to be transferred from the double ended cylinder through valve 3 for packaging and reuse in the hydrogen generation apparatus units. In addition to Eq. (13.8), there exist two other decomposition reactions that can occur in the electrolytic cell that produce carbon monoxide (CO(g)) and carbon dioxide (CO2(g)) at the anode in addition to oxygen (O2(g)), as shown in Eqs. (13.24) and (13.25).

þ þ À2 ! þ þ ð : Þ 4Na 2CO3 4NaðgÞ 2COðgÞ 2O2ðgÞ 13 24

þ þ À2 ! þ þ ð : Þ 4Na 2CO3 4NaðgÞ 2CO2ðgÞ O2ðgÞ 13 25

Electrolysis according to Eqs. (13.24) and (13.25) is undesirable and should be suppressed by identifying and using the operating conditions necessary for electrolysis according to Eq. (13.8). Table 13.12 provides the fusion and vaporization temperatures for the reactants and products in Eqs. (13.8), (13.24), and (13.25). Table 13.13 provides the enthalpies of fusion and vaporization for chemical reactants and products in Eq. (13.8). Table 13.14 provides the entropies of fusion and vaporization for chemical reactants and products in Eq. (13.8). A similar procedure as described in Sects. 13.3.1 and 13.3.2, can be used to calculate an accurate estimate of the decomposition potentials ECELL, for fused Na2CO3(l) in the novel electrolytic cell according to Eqs. (13.8), (13.24) and (13.25) at the cell operating temperature ranging between 1223.15 K (950 °C) to 1323.15 K (1050 °C), using Eqs. (13.11) and (13.12). Table 13.15 lists the standard Gibbs free energies of formation DGf , standard enthalpies of formation DHf and standard entropies S° of the compounds in Eqs. (13.8), (13.24) and (13.25)at temperature T = 298.15 K and standard atmospheric pressure of P0 = 101325 Pa. 302 A. G. Stern

Table 13.12 Fusion and vaporization temperatures at P0 = 101325 Pa

Fusion temperature Tf (K) Vaporization temperature Tb (K) Sodium Carbonate 1124.15 (Newkirk and Aliferis (Na2CO3) 1958) Sodium (Na) 370.95 ± 0.05 (Alcock et al. 1154.5 ± 4.7 (Makansi et al. 1994) 1955) aCarbon (C) (graphite) 4765 (Leider et al. 1973) 3915 (Leider et al. 1973) Carbon Monoxide 68.146 (Goodwin 1985) 81.638 (Goodwin 1985) (CO) bCarbon Dioxide 216.55 (Meyers and Van 194.635 (Meyers and Van (CO2) Dusen 1933) Dusen 1933) aFusion temperature represents the triple point at P = 110477005 Pa; Sublimes at the vaporization temperature bFusion temperature represents the triple point at P = 517984.066 Pa; Sublimes at the vaporization temperature

Table 13.13 Enthalpies of fusion and vaporization at P0 = 101325 Pa a Enthalpy of fusion DHfus (J/ Enthalpy of vaporization DHvap mol) (J/mol) Sodium Carbonate 28032.8 ± 836.8 (Janz et al. (Na2CO3) 1963) Sodium (Na) 2600 ± 25 (Alcock et al. 89240.15 (Fink and Leibowitz 1994) 1995) aEnthalpy of vaporization at the vaporization temperature provided in Table 13.12

Table 13.14 Entropies of fusion and vaporization at P0 = 101325 Pa a a Entropy of fusion DSfus Entropy of vaporization DSvap (J/KÁmol) (J/KÁmol) Sodium 24.94 (Newkirk and Aliferis Carbonate 1958; Janz et al. 1963) (Na2CO3) Sodium (Na) 7.00 (Alcock et al. 1994) 77.298 (Makansi et al. 1955; Fink and Leibowitz 1995) a Entropy of fusion and vaporization calculated using the formulas DSfus = DHfus / Tf and DSvap = DHvap / Tb, respectively

Using the tabulated standard free energies of formation DGf , in Table 13.15,it becomes possible to calculate the change in the Gibbs free energy for the electrolytic decomposition of aqueous sodium carbonate (Na2CO3(aq)) according to three possible reactions that parallel Eqs. (13.8), (13.24) and (13.25), as shown in Eqs. (13.26)–(13.28). 13 Novel Method and Molten Salt Electrolytic Cell … 303

Table 13.15 Gibbs free energy, enthalpy of formation, and entropy at T = 298.15 K and P0 = 101325 Pa ° DGf (kJ/mol) DHf (kJ/mol) S (J/KÁmol) Sodium Carbonate −1051.67 (Weast 1976; −1154.63 (Weast 1976; 70.50 (Weast 1976; (Na2CO3(aq)) Hummel et al. 2002) Hummel et al. 2002) Hummel et al. 2002) Sodium Carbonate −1047.67 (Weast −1130.94 (Weast 135.98 (Weast 1976) (Na2CO3(s)) 1976) 1976)

Sodium (Na(s)) 0 0 51.1 ± 0.3 (Alcock et al. 1994)

Carbon (C(s)) 0 0 5.69 (Weast 1976) Carbon Monoxide −137.27 (Weast 1976) −110.52 (Weast 1976) 197.91 (Weast 1976) (CO(g)) Carbon Dioxide −394.38 (Weast 1976) −393.51 (Weast 1976) 213.64 (Weast 1976) (CO2(g))

ÀÁÀÁ ÀÁ ÀÁ D ¼ Á D þ Á D þ Á D G 4ÀÁGf NaÀÁðsÞ 2 Gf CðsÞ 3 Gf O2ðgÞ À 2 Á DGf Na2CO3ðaqÞ ¼ 2103:34 kJ ð13:26Þ ÀÁÀÁ ÀÁ ÀÁ D ¼ Á D þ Á D þ Á D G 4ÀÁGf NaÀÁðsÞ 2 Gf COðgÞ 2 Gf O2ðgÞ À 2 Á DGf Na2CO3ðaqÞ ¼ 1828:80 kJ ð13:27Þ ÀÁÀÁ ÀÁ ÀÁ D ¼ Á D þ Á D þ D G 4ÀÁGf NaÀÁðsÞ 2 Gf CO2ðgÞ Gf O2ðgÞ À 2 Á DGf Na2CO3ðaqÞ ¼ 1314:58 kJ ð13:28Þ

The results of DG = 2103.34 kJ, DG = 1828.80 kJ and DG = 1314.58 kJ from Eqs. (13.26)–(13.28), respectively using aqueous Na2CO3(aq) are valid at temperature T = 298.15 K and pressure P0 = 101325 Pa. The large positive values for DG show that under these conditions, the chemical decomposition of the Na2CO3(aq) will not be spontaneous and instead, require an external input of energy. Using Eq. (13.11) with the results from Eqs. (13.26)–(13.28), it is possible to calculate the potentials required for the electrolytic cell to decompose the Na2CO3(aq), as shown in Eqs. (13.29)–(13.31).

ECELL ¼ DG=ðÀn Á FÞ¼2103340 J =ðÀ4 Á 96485:3365 C=molÞ¼À5:45 V ð13:29Þ

ECELL ¼ DG=ðÀn Á FÞ¼1828800 J =ðÀ4 Á 96485:3365 C=molÞ¼À4:74 V ð13:30Þ 304 A. G. Stern

ECELL ¼ DG=ðÀn Á FÞ¼1314580 J =ðÀ4 Á 96485:3365 C=molÞ¼À3:41 V ð13:31Þ

The potential of the electrolytic cell ECELL = −5.45 V needed to decompose Na2CO3(aq) according to Eq. (13.26), indicates that the electrolytic cell must be operated at a greater voltage than the results for both NaOH(aq) and NaCl(aq) calculated in Eqs. (13.14) and (13.22), respectively, to electrolyze Na2CO3 without CO and CO2 evolution.

13.3.3.1 Electrolysis of Na2CO3 as a Function of Temperature

To calculate the potential of the electrolytic cell needed to decompose fused or molten Na2CO3(l) at temperatures ranging between 1223.15 K (950 °C) to 1323.15 K (1050 °C) described by Eqs. (13.8), (13.24) and (13.25), it is necessary to calculate the entropy values of the reactants and products listed in Table 13.12, as well as the enthalpies of formation of the Na2CO3(l),CO(g) and CO2(g) as a function of temperature. Equation (13.12) can then be used to calculate the resulting change in the Gibbs free energy for the electrochemical decomposition reactions in Eqs. (13.8), (13.24) and (13.25) as a function of temperature. The calculation of the change in entropy as a function of temperature for the reactants and products in Eqs. (13.8), (13.24) and (13.25), is performed according to Eq. (13.15). The change in the enthalpies of formation for Na2CO3(l),CO(g) and CO2(g), is calculated according to Eq. (13.16) by evaluating the chemical reactions provided in Eqs. (13.32)–(13.34).

2NaðgÞ þ CðsÞ þ ðÞ3=2 O2ðgÞ ! Na2CO3ðlÞ ð13:32Þ

CðsÞ þ ðÞ1=2 O2ðgÞ ! COðgÞ ð13:33Þ

CðsÞ þ O2ðgÞ ! CO2ðgÞ ð13:34Þ

Table 13.16 Heat capacities at constant pressure with P0 = 101325 Pa

Temperature cp (J/KÁmol) (K) Sodium Carbonate 298.15 110.876 (Janz et al. 1963) −3 (Na2CO3) 707–1127 34.058 + (147.696 Â 10 ) ÁT (Janz et al. 1963) 1127–1210 142.214 + (44.768 Â 10−3) ÁT (Janz et al. 1963) Carbon (C) (graphite) 298.15 8.53 (McBride et al. 1963; Butland and Maddison 1973) 300 8.59 (McBride et al. 1963; Butland and Maddison 1973) 400 11.97 (McBride et al. 1963; Butland and Maddison 1973) 500 14.64 (McBride et al. 1963; Butland and Maddison 1973) 600 16.84 (McBride et al. 1963; Butland and Maddison 1973) 700 18.57 (McBride et al. 1963; Butland and Maddison 1973) 800 19.87 (McBride et al. 1963; Butland and Maddison 1973) 900 20.83 (McBride et al. 1963; Butland and Maddison 1973) 1000 21.56 (McBride et al. 1963; Butland and Maddison 1973) 1100 22.15 (McBride et al. 1963; Butland and Maddison 1973) 1200 22.68 (McBride et al. 1963; Butland and Maddison 1973) 1300 23.15 (McBride et al. 1963; Butland and Maddison 1973) 1400 23.56 (McBride et al. 1963; Butland and Maddison 1973) temperature. In Fig. 13.9, the entropies as a function of temperature are calculated for sodium carbonate (Na2CO3), sodium (Na), carbon (C), carbon monoxide (CO) and carbon dioxide (CO2). In Fig. 13.10, the enthalpy of formation as a function of temperature is calculated for sodium carbonate (Na2CO3), carbon monoxide (CO) and carbon dioxide (CO2) taking into consideration the enthalpies of fusion and vaporization given in Table 13.13. In Fig. 13.11, using the calculated entropies and enthalpies of formation as a function of temperature shown in Figs. 13.9 and 13.10, respectively, it becomes possible to calculate the change in the Gibbs free energy of the electrochemical decomposition reactions in Eqs. (13.8), (13.24) and (13.25) as a function 306 A. G. Stern

Table 13.17 Heat capacities at constant pressure with P0 = 101325 Pa

Temperature (K) cp (J/KÁmol) Carbon Monoxide (CO) 298.15 29.14 (McBride et al. 1963) 300 29.14 (McBride et al. 1963) 400 29.34 (McBride et al. 1963) 500 29.79 (McBride et al. 1963) 600 30.44 (McBride et al. 1963) 700 31.17 (McBride et al. 1963) 800 31.90 (McBride et al. 1963) 900 32.58 (McBride et al. 1963) 1000 33.18 (McBride et al. 1963) 1100 33.71 (McBride et al. 1963) 1200 34.17 (McBride et al. 1963) 1300 34.57 (McBride et al. 1963) 1400 34.91 (McBride et al. 1963)

Carbon Dioxide (CO2) 298.15 37.13 (McBride et al. 1963) 300 37.21 (McBride et al. 1963) 400 41.32 (McBride et al. 1963) 500 44.62 (McBride et al. 1963) 600 47.32 (McBride et al. 1963) 700 49.56 (McBride et al. 1963) 800 51.43 (McBride et al. 1963) 900 53.00 (McBride et al. 1963) 1000 54.31 (McBride et al. 1963) 1100 55.41 (McBride et al. 1963) 1200 56.34 (McBride et al. 1963) 1300 57.13 (McBride et al. 1963) 1400 57.80 (McBride et al. 1963)

Fig. 13.9 Entropy as a function of temperature for sodium carbonate (Na2CO3) (thin solid), sodium (Na) (thick solid), carbon (C) (thin dash), carbon monoxide (CO) (thin dash dot) and carbon dioxide (CO2) (thin dot) 13 Novel Method and Molten Salt Electrolytic Cell … 307

Fig. 13.10 Enthalpy of formation as a function of temperature for sodium carbonate (Na2CO3) (thin solid), carbon monoxide (CO) (thin dash dot) and carbon dioxide (CO2) (thin dot) of the temperature, according to Eq. (13.12), and subsequently the electrochemical potential ECELL, using Eq. (13.11). The calculations in Fig. 13.11, show that applying a potential VCELL = 3.68 V will be sufﬁcient to electrolyze Na2CO3(l) at temperatures ranging between 1223.15 K (950 °C) to 1323.15 K (1050 °C) according to Eq. (13.8), while sup- pressing the undesirable decomposition reactions in Eqs. (13.24) and (13.25) that generate CO and CO2 at the anode. The potential required is more than two times greater than the potential required for electrolytic decomposition of fused NaOH(l) according to Eq. (13.6).

Fig. 13.11 Electrochemical potential ECELL, of the electrolytic cell as a function of temperature for the reactions in Eqs. (13.8) (solid), (13.24) (dash dot) and (13.25) (dot) 308 A. G. Stern

13.4 Discussion of Results

A novel method for implementing a hydrogen fuel, sustainable, closed clean energy cycle on a large scale, based entirely on green chemistry has been described and analyzed in detail. The method uses a novel electrolytic cell meant to operate within self-contained solar powered electrolytic sodium (Na) metal production plants, that is capable of performing electrolysis on three types of molten salts individually including NaOH, NaCl, Na2CO3 or on a mixture of NaOH and NaCl, including selective electrolysis between NaOH and NaCl. The electrolytic cell operates at a high temperature range between 1223.15 K (950 °C) to 1323.15 K (1050 °C) that enables the fused salt(s) in the crucible to become completely anhydrous by evaporating any residual water (H2O) content from the salt(s) including the H2O produced at the cell anode. The high temperature of the electrolytic cell is well above the boiling point temperature of sodium (Na) metal given as Tb = 1154.5 ± 4.7 K (Makansi et al. 1955). As a result, the sodium (Na) metal is produced in a vapor state at the cathode and does not dissolve into the fused salt(s), thereby providing a direct means of separating and purifying the Na metal where otherwise, the Na produced at the cathode might dissolve into the fused salt(s) and recombine with anode products to lower the overall efficiency and production yield of Na metal by the electrolytic cell. Most significantly, the novel, hydrogen fuel, sustainable, closed clean energy cycle enables a safe, reliable, and cost effective means of scaling up sodium (Na) metal production by using either Na2CO3 or NaCl for the manufacture of Na metal in the copious quantities needed for fueling hydrogen generation systems. The electrolysis of Na2CO3 provides a practical means of avoiding excess Cl2 production that might accompany the large scale electrolysis of NaCl. The hydrogen fuel, sustainable, closed clean energy cycle can be initiated using electrolysis of Na2CO3 until sufficient Na metal exists to allow only for reprocessing of NaOH to recover the Na metal for reuse in generating H2(g) fuel, thereby avoiding Cl2 production altogether. Only if Cl2 is required by manufacturing industries, does electrolysis of NaCl have to be performed. Otherwise, the NaCl can be converted to Na2CO3 or mixed with seawater to create concentrated brine solution for use as a reactant according to Eq. (13.1) in the hydrogen generation apparatus units operating in cold climates. The novel, hydrogen fuel, sustainable, closed clean energy cycle can inhibit rising CO2 levels in the earth’s atmosphere directly as well as indirectly. Direct reduction of CO2 emissions can occur by supplanting carbon based fossil fuels in motor vehicles and in powering single family homes and light commercial establishments. Indirect reduction of CO2 emissions can occur by using the NaOH byproduct produced in the hydrogen generation apparatus units according to Eq. (13.1), to absorb or capture CO2 generated by coal fueled thermal power plants and industrial chemical processes such as the Haber–Bosch ammonia (NH3) process. The resulting Na2CO3 can subsequently undergo electrolysis according to Eq. (13.8) to recover the Na metal for reuse in hydrogen generation apparatus units. 13 Novel Method and Molten Salt Electrolytic Cell … 309

Although greater electric energy consumption is required to produce Na metal from electrolysis of Na2CO3 than from NaOH, the fused Na2CO3(l) is less corrosive than NaOH(l), and therefore, its use could potentially improve plant safety as well as extend the longevity of equipment.

13.5 Conclusion

The novel, hydrogen fuel, sustainable, closed clean energy cycle based entirely on green chemistry enables radiant solar energy to be stored cost effectively as sodium (Na) metal for generating high purity hydrogen (H2(g)) fuel on demand that is suitable for use in fuel cells without risk of contaminating sensitive catalysts that might be present. The need for the hydrogen fuel clean energy cycle fundamentally exists because the human population on earth is increasing rapidly. Some researchers believe it will plateau at 10–12 billion, while other projections reach as high as 18 billion. These are extraordinary numbers, almost unimaginable at the time of the late 18th century when carbon based fossil fuels ﬁrst began driving the Industrial Revolution in Europe. It is now well understood that energy use corre- lates closely with a nation’s economic growth and gross domestic product (GDP) and consequently, there exists enormous pressure on governments to increase rather than curtail energy use. Carbon based fossil fuels, the principal drivers of economic growth worldwide since the advent of the Industrial Revolution in Europe, cause high levels of environmental pollution and are ultimately responsible for CO2 levels in the earth’s atmosphere surpassing 0.04% (400 ppm), as conﬁrmed by measurements made in both the northern and southern hemi- spheres. Rapid population growth coupled with the rising levels of carbon dioxide (CO2) in the atmosphere, constitute a delayed yet real existential threat to humanity, especially if vegetation can no longer sustain the rate necessary for recycling the CO2 generated by human activity. The novel, hydrogen fuel, sustainable, closed clean energy cycle implemented on a large scale, will function to inhibit the adverse atmospheric effects that carbon based fossil fuel use has engendered.

References

Akella J, Vaidya SN, Kennedy GC (1969) Melting of sodium chloride at pressures to 65 kbar. Phys Rev 185(3):1135 Alcock CB, Chase MW, Itkin VP (1994) Thermodynamic properties of the group IA elements. J Phys Chem Ref Data 23(3):385–497 Ambrose D, Hall DJ, Lee DA, Lewis GB, Mash CJ (1979) The vapour pressure of chlorine. J Chem Thermodyn 11(11): 1089–1094 Amendola SC, Sharp-Goldman SL, Saleem Janjua M, Spencer NC, Kelly MT, Petillo PJ, Binder M (2000) A safe, portable, hydrogen gas generator using aqueous borohydride solution and Ru catalyst. Int J Hydrogen Energy 25(10):969–975 310 A. G. Stern

Arblaster JW (2013) The thermodynamic properties of chlorine condensed phases. J Chem Thermodyn 56:12–14 Archer DG (1992) Thermodynamic properties of the NaCl + H2O system II. Thermodynamic properties of NaCl(aq), NaCl Á 2H2O(cr), and phase equilibria. J Phys Chem Ref Data 21 (4):793–829 Baxter PJ, Kapila M, Mfonfu D (1989) Lake Nyos disaster, Cameroon, 1986: the medical effects of large scale emission of carbon dioxide? BMJ 298(6685):1437–1441 Bergthorson JM, Yavor Y, Palecka J, Georges W, Soo M, Vickery J, Goroshin S, Frost DL, Higgins AJ (2017) Metal-water combustion for clean propulsion and power generation. Appl Energy 186:13–27 Bessarabov D, Wang H, Li H, Zhao N (eds) (2016) PEM electrolysis for hydrogen production— principles and applications. CRC Press-Taylor & Francis Group, Boca Raton, Florida Bodnar RJ (1993) Revised equation and table for determining the freezing point depression of H2O-NaCl solutions. Geochim Cosmochim Acta 57(3):683–684 Brayard F (2000) An early report by Kurt Gerstein. Bulletin du Centre de recherche français à Jérusalem 6:157–176 Buggeln M (2015) Experten der Vernichtung: Das T4-Reinhard-Netzwerk in den Lagern Belzec, Sobibor und Treblinka. German History 33(2):327–329 Butland ATD, Maddison RJ (1973) The speciﬁc heat of graphite: an evaluation of measurements. J Nucl Mater 49(1):45–56 Carmo M, Fritz DL, Mergel J, Stolten D (2013) A comprehensive review on PEM water electrolysis. Int J Hydrogen Energy 38(12):4901–4934 Castner HY (1891) Process of manufacturing sodium and potassium, U.S. Patent 452,030 Cohen JE (1995) Population growth and earth’s human carrying capacity. Science 269(5222):341– 346 Cohen JE (2003) Human population: the next half century. Science 302(5648):1172–1175 Davy H (1808) The Bakerian lecture: on some new phenomena of chemical changes produced by electricity, particularly the decomposition of the ﬁxed alkalies, and the exhibition of the new substances which constitute their bases; and on the general nature of alkaline bodies. Philos Trans R Soc Lond 98:1–44 Dlugokencky EJ, Hall BD, Crotwell MJ, Montzka SA, Dutton G, Mühle J, Elkins JW (2016) Atmospheric composition. In: Blunden J, Arndt DS (eds) [State of the climate in 2015], Bulletin of the American meteorological society, vol 97, no 8, pp 1–241 Doherty BT, Kester DR (1974) Freezing-point of seawater. J Mar Res 32(2):285–300 Downs JC (1924) Electrolytic process and cell, U.S. Patent 1,501,756 Dworkin AS, Bredig MA (1960) The heat of fusion of the alkali metal halides. J Phys Chem 64 (2):269–272 Feistel R, Wagner W (2005) High-pressure thermodynamic Gibbs function of ice and sea ice. J Mar Res 63(1):95–139 Feistel R, Wagner W (2006) A New Equation of State for H2O Ice lh. J Phys Chem Ref Data 35 (2):1021–1047 Fink JK, Leibowitz L (1995) “Thermodynamic and transport properties of sodium liquid and vapor”, ANL/RE-95/2. Argonne National Laboratory, Argonne, IL Giauque WF, Powell TM (1939) Chlorine. The heat capacity, vapor pressure, heats of fusion and vaporization, and entropy. J Am Chem Soc 61(8):1970–1974 Gibbs JW (1878) On the equilibrium of heterogeneous substances. Am J Sci 16(96):441–458 Goodwin RD (1985) Carbon monoxide thermophysical properties from 68 to 1000 K at pressures to 100 MPa. J Phys Chem Ref Data 14(4):849–932 Guais A, Brand G, Jacquot L, Karrer M, Dukan S, Grevillot G, Molina TJ, Bonte J, Regnier M, Schwartz L (2011) Toxicity of carbon dioxide: a review. Chem Res Toxicol 24(12):2061–2070 Gurvich LV, Bergman GA, Gorokhov LN, Iorish VS, Leonidov VYa, Yungman VS (1996) Thermodynamic properties of alkali metal hydroxides. Part 1. Lithium and sodium hydroxides. J Phys Chem Ref Data 25(4): 1211–1276 13 Novel Method and Molten Salt Electrolytic Cell … 311

Hall DL, Sterner SM, Bodnar RJ (1988) Freezing point depression of NaCl-KCl-H2O solutions. Econ Geol 83(1):197–202 Hilsenrath J, Beckett CW, Benedict WS, Fano L, Hoge HJ, Masi JF, Nuttall RL, Touloukian YS, Woolley HW (1955) Tables of thermal properties of gases. US Department of Commerce, National Bureau of Standards Circular vol 564, pp 267–296 Hummel W, Berner U, Curti E, Pearson FJ, Thoenen T (2002) Nagra/PSI chemical thermodynamic data base 01/01, Technical Report 02-16 Janz GJ, Neuenschwander E, Kelly FJ (1963) High-temperature heat content and related properties for Li2CO3,Na2CO3,K2CO3, and the ternary eutectic mixture. Trans Faraday Soc 59:841–845 Klanchar M, Wintrode BD, Phillips JA (1997) Lithium-water reaction chemistry at elevated temperature. Energy Fuels 11(4):931–935 Kling GW, Clark MA, Compton HR, Devine JD, Evans WC, Humphrey AM, Koenigsberg EJ, Lockwood JP, Tuttle ML, Wagner GN (1987) The 1986 Lake Nyos gas disaster in Cameroon, West Africa. Science 236:169–176 Kojima Y, Suzuki K, Fukumoto K, Sasaki M, Yamamoto T, Kawai Y, Hayashi H (2002) Hydrogen generation using sodium borohydride solution and metal catalyst coated on metal oxide. Int J Hydrogen Energy 27(10):1029–1034 Kojima Y, Kawai Y, Nakanishi H, Matsumoto S (2004a) Compressed hydrogen generation using chemical hydride. J Power Sources 135(1–2):36–41 Kojima Y, Suzuki K, Fukumoto K, Kawai Y, Kimbara M, Nakanishi H, Matsumoto S (2004b) Development of 10 kW-scale hydrogen generator using chemical hydride. J Power Sources 125(1):22–26 Kojima Y, Kawai Y, Kimbara M, Nakanishi H, Matsumoto S (2004c) Hydrogen generation by hydrolysis reaction of lithium borohydride. Int J Hydrogen Energy 29(12):1213–1217 Kong VCY, Kirk DW, Foulkes FR, Hinatsu JT (2003) Development of hydrogen storage for fuel cell generators II: utilization of calcium hydride and lithium hydride. Int J Hydrogen Energy 28 (2):205–214 Leadbetter AJ, Settatree GR (1969) Anharmonic effects in the thermodynamic properties of solids IV. The heat capacities of NaCl, KCl, KBr between 30 and 500 °C. J Phys C Solid State Phys 2 (3):385–392 Leider HR, Krikorian OH, Young DA (1973) Thermodynamic properties of carbon up to the critical point. Carbon 11(5):555–563 Lorenz R, Winzer R (1929) Die Löslichkeit von Natrium und Calcium in ihren Chloriden und Chloridgemischen. Zeitschrift für anorganische und allgemeine Chemie 183(1):121–126 Lutz W, Samir KC (2010) Dimensions of global population projections: what do we know about future population trends and structures? Philos Trans R Soc London B Biol Sci 365 (1554):2779–2791 Makansi MM, Muendel CH, Selke WA (1955) Determination of the vapor pressure of sodium. J Phys Chem 59(1):40–42 Marini S, Salvi P, Nelli P, Pesenti R, Villa M, Berrettoni M, Zangari G, Kiros Y (2012) Advanced alkaline water electrolysis. Electrochim Acta 82:384–391 McBride BJ, Heimel S, Ehlers JG, Gordon S (1963) “Thermodynamic properties to 6000° K for 210 substances involving the ﬁrst 18 elements”, NASA SP-3001. National Aeronautics and Space Administration, Washington, DC Meyers CH, Van Dusen MS (1933) The vapor pressure of liquid and solid carbon dioxide. Bur Stand J Res 10(3):381–412 Millero FJ, Feistel R, Wright DG, McDougall TJ (2008) The composition of standard seawater and the deﬁnition of the reference-composition salinity scale. Deep-Sea Res Part I Oceanogr Res Papers 55(1):50–72 Muir SS, Yao X (2011) Progress in sodium borohydride as a hydrogen storage material: development of hydrolysis catalysts and reaction systems. Int J Hydrogen Energy 36 (10):5983–5997 Newkirk AE, Aliferis I (1958) Drying and decomposition of sodium carbonate. Anal Chem 30 (5):982–984 312 A. G. Stern

Redkin AA, Zaikov YP, Korzun IV, Reznitskikh OG, Yaroslavtseva TV, Kumkov SI (2015) Heat capacity of molten halides. J Phys Chem B 119(2):509–512 Satish U, Mendell MJ, Shekhar K, Hotchi T, Sullivan D, Streufert S, Fisk WJ (2012) Is CO2 an indoor pollutant? Direct effects of low-to-moderate CO2 concentrations on human decision-making performance. Environ Health Perspect 120(12):1671–1677 Schlesinger HI, Brown HC, Finholt AE, Gilbreath JR, Hoekstra HR, Hyde EK (1953) Sodium borohydride, its hydrolysis and its use as a reducing agent and in the generation of hydrogen. J Am Chem Soc 75(1):215–219 Selman JR, Maru HC (1981) Physical chemistry and electrochemistry of alkali carbonate melts, with special reference to the molten-carbonate fuel cell. In: Mamantov G, Braunstein J (eds) Advances in molten salt chemistry, volume 4. Plenum Press, New York, New York, pp 159–390 Sigurdsson H, Devine JD, Tchua FM, Presser FM, Pringle MKW, Evans WC (1987) Origin of the lethal gas burst from Lake Monoun, Cameroun. J Volcanol Geoth Res 31(1–2):1–16 Steinhauser G (2008) Cleaner production in the solvay process: general strategies and recent developments. J Clean Prod 16(7):833–841 Stern AG (2015) Design of an efﬁcient, high purity hydrogen generation apparatus and method for a sustainable, closed clean energy cycle. Int J Hydrogen Energy 40(32):9885–9906 Stern AG (2017) Scalable, self-contained sodium metal production plant for a hydrogen fuel clean energy cycle. In: Nikolic AB, Janda ZS, (eds) Recent improvements of power plants management and technology. InTech Publisher, Vienna, pp 145–189. ISBN: 978-953-51-3357-5 Stern AG (2018) A new sustainable hydrogen clean energy paradigm. Int J Hydrogen Energy 43 (9):4244–4255 Stewart RB, Jacobsen RT, Wagner W (1991) Thermodynamic properties of oxygen from the triple point to 300 K with pressures to 80 MPa. J Phys Chem Ref Data 20(5):917–1021 Von Hevesy G (1909) Über die schmelzelektrolytische Abscheidung der Alkalimetalle aus Ätzalkalien und die Löslichkeit dieser Metalle in der Schmelze. Zeitschrift für Elektrochemie 15(15):529–536 Wagner W, Prub A (2002) The IAPWS formulation 1995 for the thermodynamic properties of ordinary water substance for general and scientiﬁc use. J Phys Chem Ref Data 31(2):387–535 Wartenberg HV, Albrecht P (1921) Die Dampfdrucke einiger Salze. Zeitschrift für Elektrochemie 27:162–167 Weast RC (1976–1977) Handbook of chemistry and physics. CRC Press, Cleveland, Ohio, C415, D67–D78, D79–D84, D141–D146 Xu J, Froment GF (1989) Methane steam reforming, methanation and water-gas shift: I. Intrinsic kinetics. AIChE J 35(1):88–96 Zeng K, Zhang D (2010) Recent progress in alkaline water electrolysis for hydrogen production and applications. Prog Energy Combust Sci 36(3):307–326 Zhang JZ, Li J, Li Y, Zhao Y (2014) Hydrogen generation, storage and utilization. Wiley, Hoboken, New Jersey, pp 24–46, 51–69 Chapter 14 Renewable Energy Storage and Its Application for Desalination

Muhammad Wakil Shahzad, Muhammad Burhan and Kim Choon Ng

Abstract The economic development has serious impact on the nexus between water, energy, and environment. This impact is even more severe in Non-Organization for Economic Cooperation and Development (non-OECD) countries due to improper resource management. It is predicted that energy demand will increase by more than 71% in non-OECD as compared to 18% in developed countries by 2040. In Gulf Cooperation Council countries, water and power sector consume almost half of primary energy produced. In the past, many studies were focused on renewable energies based on desalination processes to accommodate ﬁvefold increase in demand by 2050 but they were not commercialized due to intermittent nature of renewable energy such as solar and wind. We proposed highly efﬁcient energy storage material, magnesium oxide (MgO), system integrated with innovative hybrid desalination cycle for future sustainable water supplies. The condensation of Mg(OH)2 dehydration vapor during day operation with concentrated solar energy and exothermic hydration of MgO at night can produce 24 h thermal energy without any interruption. It was showed that Mg(OH)2 dehydration vapor condensation produces 120 °C and MgO hydration exothermic reaction produces 140 °C heat during day and night operation, respectively, corresponding to energy storage of 81 kJ/mol and 41 kJ/mol. The produced energy can be utilized to operate desalination cycle to reduce CO2 emission and to achieve COP21 goal. The proposed hybrid desalination cycle is successfully demonstrated by pilot experiments at KAUST. It was showed that MgO + MEDAD cycle can achieve performance over UPR = 200, one of the highest reported ever.

Keywords Thermal desalination Á Adsorption cycle Á Hybrid cycle Renewable energy Á Solar energy Á Energy storage material

M. W. Shahzad (&) Á M. Burhan Á K. C. Ng Water Desalination and Reuse Centre (WDRC), Biological and Environmental Science & Engineering (BESE), King Abdullah University of Science and Technology, Thuwal, Saudi Arabia e-mail: [email protected]

© Springer Nature Singapore Pte Ltd. 2019 313 E. Motoasca et al. (eds.), Energy Sustainability in Built and Urban Environments, Energy, Environment, and Sustainability, https://doi.org/10.1007/978-981-13-3284-5_14 314 M. W. Shahzad et al.

Nomenclature OECD Organization for Economic Cooperation and Development COP Conference of Parties GCC Gulf Cooperation Council EAD Environment Agency Abu Dhabi MENA Middle East and North Africa MED Multi-effect Desalination MSF Multistage Flash SWRO Seawater Reverse Osmosis MVC Mechanical Vapor Compression AD Adsorption ED Electrodialysis TES Thermal Energy Storage CSP Concentrated Solar Photovoltaic TBT Top Brine Temperature LBT Lower Brine Temperature LPM Liter per Minute UPR Universal Performance Ratio TL Thermodynamic Limit

14.1 Introduction

Increased water demand and the degradation of water sources threaten not only the fast economic development activities but also the public health in many developing countries. Particularly, GCC countries are most affected region because of rapid population and development growth and poor water resources management even with limited water resources. In GCC region, per capita renewable water availability is the lowest in the world and they housed 5% of world population. To meet water demand, all GCC countries heavily rely on desalination processes. The GCC countries started desalination in 1950 to meet their demand and to maintain groundwater resources’ sustainability and today they hold around 75% share of world desalination capacities. The change in climate is expected to make this region more water scarier due to high evaporation rate and reduced average rainfall. In addition, the population growth trend coupled with anticipated climate change will pressurize the water situation even more. It is estimated that by 2050, GCC countries will only able to satisfy 23% of demand that will yield 26 billion cubic meters (bcm) per year water shortage, 77% of consumption (Parmigiani 2015). Huge desalination capacities are expected to be installed in future to fulﬁll the water demand. Unfortunately, the conventional desalination process is not only energy intensive but also environmentally unfriendly. Today’s best desalination processes consume 7–10 kWh primary energy to produce one cubic meter of water and emit 14 Renewable Energy Storage and Its Application for Desalination 315 around 3–4kgofCO2 emission (Munawwar and Ghedira 2014). The amount of energy required for seawater desalination processes is almost 10 times higher than the river and lake water treatment. This high energy consumption not only results in higher water and associated processes cost but also severe environmental impact. Therefore, an improved desalination process and implementation of alternate energy resources are important for future sustainability. Since renewable energies have no impact on environment, desalination processes based on these technologies will be more sustainable for future water supplies. In addition, the overall cost of water production will be reduced to very low due to free availability of these energy sources. Renewable energy-driven desalination process is in R&D phase since long and few of them were able to get place in practical application but most of them were not able to commercialize due to large capital cost and intermittent nature of renewable energy. The Sunbelt region, GCC countries, have highest solar irradiance and also available most of the time of the year that can be harnessed to produce potable water through desalination processes in sustainable manner. Figure 14.1 shows the solar irradiance in GCC region (Renewable Energy Market analysis 2016). Solar energy is an attractive source because ﬁrst its has low operational and capital cost and in addition, it is environmentally friendly. Solar energy-driven desalination processes are heavily investigated by many researchers for their commercial application possibilities. Some processes are still under research and development phase, and some technologies are implemented at certain locations for commercial application. Figure 14.2 shows the status of desalination processes operating with solar energy. The application chart showed that solar energy-based thermally driven desalination processes are implementing at large scale to produce freshwater followed by photovoltaic-driven membrane processes (Shahzad 2017a). Even though solar energy is promising for sustainable water production, its intermittent availability is the main bottleneck in its major application. Currently, over 130 desalination plants powered by renewable processes contributing only 1%

Fig. 14.1 Solar energy availability in Middle East and North Africa (MENA) region 316 M. W. Shahzad et al.

Solar Still Photovoltaic Solar/CSP <0.1m3/day, SWRO MED 3 3 1.3-6.5 $/m3 <100m /d, >5000m /d, 11.7-15.6 $/m3 2.0-2.5 $/m3

Application Wind RO 3 3 Solar multi-effect humidification 50-2000m /d, 2.0-5.2 $/m 1-100m3/d, 2.6-6.5 $/m3

Geothermal MED 50-1000m3/day Solar membrane distillation 0.15-10m3/d, Advanced R&D R&D Advanced 10.4-19.5 $/m3 Wind MVC <100m3/d, 5.2-7.8 $/m3 Photovoltaic EDR <100m3/d, 10 $/m3

Wind ED Solar organic Basic Research Wave RO SWRO 1000-3000m3/d, >0.1m3/d 0.7-1.2 $/m3, Basic Research

L / day m3 / day 100x m3/day 1000x m3/day

Fig. 14.2 Renewable energy-driven desalination technologies and their status of application with maximum design size

Fig. 14.3 Share of different desalination processes operating with renewable energy sources 14 Renewable Energy Storage and Its Application for Desalination 317 to total world desalination capacities. The share of different processes is shown in Fig. 14.3 (Shatat and Riffat 2014). To increase solar-driven desalination capacity, it requires energy storage for 24 h operation.

14.2 Solar Thermal Energy Storage Options

Two main features of solar energy, almost free and endless, make it more attractive as compared to other renewables. However, its cyclic nature is the main hindrance in its commercial applications. To overcome this limitation, scientists proposed solar thermal energy storage (STES) system that can store solar energy as a thermal heat during daytime and release when solar radiations are weak or not available. The necessary electricity for equipment operation can be supplied with concentrated solar photovoltaic (CSP) system equipment with STES system. The commercial applicability of STES system depends on its efficiency and cost-effectiveness. The STES system can be divided into three classes based on the method of heat storage. These three classes are the sensible heat storage, the latent heat storage, and thermochemical heat storage. The efficiency and cost-effectiveness of STES system are based on five parameters such as (i) temperature for storage, (ii) period for storage, (iii) storage material availability, (iv) material storage capacity, and (v) operation. The thermochemical heat storage has many advantages such as (i) 5–10 times higher energy density as compared to other two STES systems and (ii) minimum losses during storage and transportation. Because of these two properties, these systems can be utilized to store long-term period energy. However, thermochemical heat storage system is still not able to commercialized at large scale and this is one of the reasons that there is no trustable data available for application design. The designing of thermochemical heat storage system needs careful selection of chemical reaction and detailed investigation of operational parameters such as pressure, temperature, kinetics, and reversibility. The main classifications of reactions for thermochemical heat storage system are presented in Fig. 14.4 (Shahzad 2017a). A chemical heat pump (CHP) operated with magnesium oxide and water pair (MgO/H2O) is considered due to high energy storage capacity and simple operation. The proposed CHP operates on reversible chemical reaction between

C1 C2 C3 C4 C5 C6 Metal Carbonates Redox Ammonia Organic Hydrides Hydroxydes

CH4/H2O MgH PbCO3 Mg(OH)2 BaO2 NH4HSO4 2 CH4/CO2 CaH CaCO3 Ca(OH)2 Co3O4 NH3 2 SO3

Fig. 14.4 Main classiﬁcations of reactions for thermochemical heat storage system 318 M. W. Shahzad et al.

Fig. 14.5 Energy density of different thermal energy storage materials at different operating temperatures

ð Þ MgO and Mg OH 2 based on the following equilibria. Magnesium oxide has highest energy storage density as compared to many other energy storage materials as shown in Fig. 14.5 (Shatat and Riffat 2014; Pardo et al. 2014; Kato et al. 1997, 1998).

ðÞþ ðÞ$ ð Þ ðÞ D ¼À : ½ MgO s H2Og Mg OH 2 s H1 81 02 KJ/mol

H2OgðÞ$H2OlðÞ DH1 ¼À40:66½ KJ/mol

The operation of proposed CHP system is presented in Fig. 14.6 (Kato et al. 1996, 1999, 2009). The two major components are (i) chemical reactor packed with magnesium oxide and (ii) an evaporator ﬁlled with water. One complete cycle of operation consists of heat storage mode and heat release mode. In heat storage mode, the fully hydrated magnesium hydroxide (Mg(OH)2)is exposed to the concentrated sun at Td to supply heat (Qd) for dehydration process. The vapors produced during dehydration process are condensed at TC in the condenser, and the condensation heat (Qd) is utilized for desalination cycle at daytime during energy storage mode. In heat release mode, hydrated magnesium oxide is exposed to water evaporator to extract vapor for hydration. Since the adsorption is an exothermic reaction, hydration heat (Qh)atTh is extracted and utilized during night operation and heat release mode. It can be noticed that there are no major moving parts that are involved in operation, and this is the major advantage of proposed cycle. In addition, it is 14 Renewable Energy Storage and Its Application for Desalination 319

Fig. 14.6 Energy storage principle of MgO/H2O heat pump environmentally friendly and also economic because MgO is easily available. The proposed CHP can store energy at over 300 °C and deliver over 100 °C for 24 h operation. This heat can be utilized for the operation of proposed cycles for cooling and desalination. The detail of proposed cycles, adsorption (AD) cycle for cooling, and hybrid multi-effect desalination and adsorption cycle (MEDAD) for desalination are presented in the following sections.

14.3 STES System and Desalination Cycle

The proposed system has three major subsystems, namely the AD subsystem, the MED subsystem, and STES subsystem. The detail of AD subsystem, hybrid MEDAD subsystem, and MgO + MEDAD is presented in the following sections.

14.3.1 AD Subsystem

Low-temperature-operated adsorption (AD) subsystem is an affordable solution to produce freshwater and cooling using solar thermal energy. Plethora literature is available on AD cycle investigation for their working pair selection and performance improvement. The basic AD cycle consists of adsorbent beds, evaporator, condenser, and liquid circulation pumps. The process schematic is shown in Fig. 14.7. The two main processes for AD cycle are adsorption–evaporation and desorption–condensation. In adsorption–evaporation process, the dried adsorbent is exposed to the evaporator to assist evaporation due to adsorbent high afﬁnity for water vapors, 320 M. W. Shahzad et al.

Fig. 14.7 Basic adsorption subsystem ﬂow schematic since adsorption is an exothermic process that increases the adsorbent temperature and reduces its adsorption capacity. To maintain adsorption process, coolant is circulated through heat exchanger tubes to extract the heat of adsorption. The evaporator temperature is maintained by circulating the chilled water through the tubes. The AD evaporator can operate at as low as 5 °C to produce cooling effect when it is operated as a chiller. The adsorbent can adsorb vapors and maintain evaporation process until it gets fully saturated. Once adsorbent is fully saturated, it switches process to regeneration to prepare for next operation. In desorption–condensation process, low-temperature heat source from 55 to 85 °C is supplied to the reactor bed to regenerate the adsorbent. The regenerated vapors are then condensed in the condenser to produce distillate. When it is operating as a chiller, the refrigerant is reﬂuxed back to the evaporator as a closed loop. To overcome cyclic operation limitations of conventional AD cycle, multi-bed system was proposed for continuous operation. The number of beds depends on the size of AD system as per cooling or desalination requirements. In multi-bed scheme, they operate as a pair for adsorption and desorption processes. The automatic switching method is applied to switch their duties from adsorption to desorption process to complete cycle. The time for adsorption/desorption process depends on quantity of adsorbent packed in a single bed. It can be noticed that the major moving parts are just automatic valves, and circulation pumps that make AD cycle very robust and very less maintenance are required. It can produce two useful effects, cooling and desalinated water. 14 Renewable Energy Storage and Its Application for Desalination 321

14.3.2 Hybrid MEDAD Desalination Subsystem

In conventional multi-effect desalination (MED) subsystem, the heat source is supplied to the first steam generator and seawater is sprayed on to the tube surfaces. The saturation pressure is controlled to maintain evaporation under vacuum conditions at 65 °C. The vapor produced in the first stage is channeled to the tubes of second stage to reutilize the heat of condensation. The produced vapors are then condensed by transferring the heat condensation to the seawater sprayed that help to evaporate the part of the feed. The condensed vapor produces freshwater that is collected from each stage. This process of evaporation and condensation of feed water and vapors continues until the last stage of the system. The last-stage vapors are condensed in the final condenser by circulating the seawater. It can be noticed that the overall operational range for conventional MED system is from heat source temperature 65 °C to the last-stage condenser temperature 40 °C. These operational limitations also limit the performance of thermally driven desalination processes. The performance of desalination processes can be improved by overcoming their operational limitations. Some studies showed that heat source temperature can be increased to 120 °C by introducing proper pretreatment processes to overcome the scaling and fouling chances. Nanofiltration (NF) is one of the pretreatment processes that can help to remove soft scaling components from feed water to prevent scaling and fouling chances at high-temperature operation. On the other hand, the proposed hybridization of AD cycle with conventional MED system can help to overcome last condenser temperature limitations and operation can be extended to as low as 5 °C. The detailed schematic of MEDAD

Fig. 14.8 Hybrid MED + AD desalination subsystem ﬂow schematic 322 M. W. Shahzad et al. hybrid cycle is shown in Fig. 14.8 (Shahzad 2013; Shahzad and Ng 2016, 2017; Shahzad et al. 2014, 2015, 2017b, 2016a, b; Thu and Kim 2015; Ng et al. 2015). In hybrid cycle, the MED last-stage condenser is bypassed and vapors are directly adsorbed on the adsorbent of AD cycle. The high afﬁnity of adsorbent toward water vapor reduced saturation pressure and hence temperature. This hybrid cycle extends the temperature range from typical 40 to 5 °C and it helps to insert more number of stags. The more number of stages will help to increase the number of recoveries from provided input, and hence performance will improve. It was reported that hybrid MEDAD cycle can boost the water production to twofold as compared to conventional MED system at same heat source temperature. It can be observed that the trihybrid system NF + MED + AD can be the best performance system if operated with renewable energy for future sustainability. The proposed trihybrid cycle can be operated with STES system to produce freshwater in 24 h. The detailed system is presented in the following sections.

14.3.3 Proposed MgO + MEDAD Desalination Subsystem

In this proposed subsystem, MgO is coated on tubes and module is sized as per energy storage requirement. This energy storage tube module is connected to a seawater evaporator and a condenser. During day operation, to dehydrate Mg(OH)2, solar rays are concentrated to the module at 300 °C and desorbed vapors are directed toward the condenser for condensation. The heat produced during

Fig. 14.9 Proposed MgO + MED + AD desalination subsystem flow schematic 14 Renewable Energy Storage and Its Application for Desalination 323 condensation is recovered through hot water closed circulation loop to supply heat to MEDAD cycle steam generator or first stage. This process continues until solar energy is available throughout the day. After sunset, during night operation, the dried MgO tubes are exposed to seawater evaporator. The hydration of MgO is an exothermal process, and this heat is recovered from the top head of tubes (acting like heat pipe) and supplied to the first stage of MEDAD cycle. The exothermic heat recovery will not only help to maintain adsorption process but will also help to operate MEDAD cycle at night. The detailed flow schematic of the proposed system is shown in Fig. 14.9. To verify the concept, a pilot was installed at King Abdullah University of Science and Technology (KAUST) and tested successfully. The detail of pilot is presented in the following section.

14.4 MEDAD Desalination Pilot Experimentation

To investigate MEDAD operation, a four-effect pilot was designed and build in KAUST as shown in Fig. 14.10 (a, b). The pilot capacity is 4 m3/day as a standalone MED and around 8 m3/day when operated with combination of Ad cycle as a MEDAD hybrid cycle. The MED pilot is operating with parallel feed system supplied from Red Sea after minimal pretreatment. Due to budget constraint, the heat input was simulated via solar collectors and boiler. Figure 14.11 shows the temperature profiles of all MED effects operating at 50 °C. The inter-effect temperature difference was noted as 2–3 °C, same as in commercial MED plants. The condenser cooling water is supplied from Red Sea before supplying as a feed and average temperature was observed at 33 °C. After 72 h successful operation, MED system was hybridized to AD cycle to demonstrate as MEDAD cycle. The experiment was continued for another 72 h and temperature trends are plotted in Fig. 14.12. It can be seen that after hybridization, the inter-effect temperature difference was increased to 4–5 °C and last effect was operating at 25 °C, below feed water temperature, as compared to conventional MED system operating at 42 °C. The MED operation below ambient temperature is demonstrated for the first time in the desalination history. The successful operation will provide opportunity of MED system to operate from top temperature 50 °C to last effect temperature to as low as 5 °C as AD cycle can help to pull temperature to 5–7 °C due to silica gel affinity for water vapors. This excellent thermodynamic synergy between MED and AD will not only help to extend conventional MED system operational rage for an additional recoverie but will also boost water production to almost twofold as presented in Fig. 14.13. The twofold water production boost due to hybridization processes can be noticed clearly from the production profiles at same heat supplied conditions. 324 M. W. Shahzad et al.

Fig. 14.10 a MEDAD hybrid pilot installed at KAUST. b MEDAD pilot control (HMI) system 14 Renewable Energy Storage and Its Application for Desalination 325

o Inter-stage temperature differential 2-3 C

Fig. 14.11 Four stages MED and condenser temperature trends at heat source 50 °C for 72 h operation

o Inter-stage temperature differential 4-5 C

Fig. 14.12 Temperature proﬁles of MEDAD hybrid cycle effects and condenser at heat source 50 °C for 72 h operation 326 M. W. Shahzad et al.

2.1 LPM 4.1 LPM

Fig. 14.13 Water production proﬁles of conventional MED and MEDAD hybrid cycle at heat source 50 °C

14.5 Desalination Processes Performance Evaluation

Currently, different methods are employed to evaluate diverse desalination processes. For example, thermally driven desalination processes are evaluated by performance ratio (PR) and pressure-driven processes are based on specific energy consumption (kWh/m3). In industry, operators sometimes use specific energy consumption method for both processes by considering the different grades of energies as same such as electricity and low-pressure steam as shown in Eq. 14.1. This misconception of considering all energies as same grade provides distorted evaluation for different desalination processes. Since derived energies (electricity and steam) involved their generation processes efficiencies, they must be converted to primary energy before considering for evaluation processes. 0 1 Equivalent heat of evaporation B C B of distillate production C PR ¼ @ A Energy input no kJ 2326 kg ffi ÂÃÈÉÈÉÈÉ ð14:1Þ : kWhelec þ kWhther þ kWhRenewable 3 6x m3 m3 m3

We proposed universal performance ratio (UPR) method for all kind of desalination processes of evaluating a common platform because it is based on primary energy. All derived energies should be converted to primary energy by considering their conversion factors and then can be utilized to calculate the performance as shown in Eq. 14.2. The proposed method can be utilized for any processes efﬁ- ciency calculation and it can be compared with any other technology because they are all based on primary energy basis. 4RnwbeEeg trg n t plcto o eaiain327 Desalination for Application Its and Storage Energy Renewable 14

Table 14.1 Universal performance ratio comparison of all desalination processes Desalination Electrical Thermal energy Conversion Conversion factor Primary UPR UPR percentage of TL technology energy consumption factor for for thermal energy energy (UPR at TL = 828) (%) consumption electricity

(kWhelec ) (kWhther ) (57.2%) (3.4%) CF2 = 29.4 (kWhpe) CF1 = 0.572 SWRO 3.85 NA 0.572 29.4 6.75 95.8 11.6 MED 1.50 61.0 4.67 138.4 16.7 MSF 2.50 72.0 8.37 77.2 9.3 MEDAD 1.80 30.5 4.19 155.0 18.7 hi 2326 kJ/kg : : ÃTL ¼ ¼ 828; 0 78 kWh ¼ 2 8 kJ 2:8 kJ/kg m2 kg 328 M. W. Shahzad et al. no kJ 2326 kg UPR ﬃ ÂÃÈÉ ÈÉ ÈÉ ð14:2Þ : kWhelec þ kWhther þ kWhRenewable 3 6x m3 CF1 m3 CF2 m3 CF3 CF ¼ conversion factor 1 ¼ electrical; 2 ¼ thermal and 3 ¼ renewable

In a cogeneration-based desalination plant, detailed exergetic analysis showed that gas turbine cycle consumes around 73.17% on input fuel exergy. The heat recovery steam generator helps to recover 26.83% exergy from exhaust gases to produce steam for steam cycle and desalination processes. The Rankine cycle consumes around 23.43% and desalination processes only utilize 3.4% on input fuel exergy. Considering these proportions, the conversion factors are calculated to convert derived energies to primary energy to compute the UPR for desalination processes as summarized in Table 14.1. It can be seen that MED processes performance is the highest because it has the best thermodynamic synergy when combined with power plants. Most importantly, all desalination processes are operating only at 10–15% of thermodynamic limit. This shows that the current desalination processes are not sustainable and their efﬁciency needs to improve to achieve 25–30% of thermodynamic limit to achieve future sustainability goals. High-efﬁciency desalination processes integrated with renewable energy sources can be the best choice for future water supplies (Shahzad 2018;Ng2017). It can be seen that even with paid thermal energy, MEDAD hybrid cycle has highest UPR as compared to other processes. In case of MgO + MEDAD, the thermal energy will be free on small expense of electrical to drive circulation pumps that will boost UPR to over 200.

14.6 Conclusions

Thermally driven desalination system and solar thermal energy storage system have been proposed for 24 h operation. It was shown that renewable energies were not implemented at commercial scale due to intermittent nature. An efficient and reliable storage system can overcome this limitation, and these free energy sources can be applied for sustainable production of desalinated water. Magnesium oxide-based CHP energy storage system is proposed to supply thermal energy to hybrid MED + AD desalination system to produce desalinated water to fulfill future water demand. The desalination processes operating with renewable energy sources will not help to reduce environmental impact but also overall water production cost. These sustainable energy sources will help to fill the gap between future water demand and supply in sustainable manner. The proposed system performance is successfully demonstrated at KAUST and it is estimated to have highest UPR, over 200. These innovative solutions will help to save energy and protect environment. 14 Renewable Energy Storage and Its Application for Desalination 329

References

Kato Y, Yamashita N, Kobayashi K, Yoshizawa Y (1996) Kinetic study of the hydration of magnesium oxide for a chemical heat pump. Appl Therm Eng 16:853–862 Kato Y, Harada N, Yoshizawa Y (1997) Kinetic feasibility of a chemical heat pump for heat utilization of high-temperature processes. Appl Therm Eng 19:239–254 Kato Y, Saku D, Harada N, Yoshizawa Y (1998) Utilization of high temperature heat from nuclear reactor using inorganic chemical heat pump. Prog Nucl Energy 32:563–570 Kato Y, Nakahata J, Yoshizawa Y (1999) Durability characteristics of the hydration magnesium oxide under repetitive reaction. J Mater Sci 34:475–480 Kato Y, Takahashi F, Sekiguchi T, Ryu J (2009) Study on medium temperature chemical heat storage using mixed hydroxides. Int J Refrig 32:661–666 Munawwar S, Ghedira H (2014) A review of renewable energy and solar industry growth in the GCC region. Energy Procedia 57:3191–3202 Ng KC, Thu K, Oh SJ, Ang L, Shahzad MW, Ismail AB (2015) Recent developments in thermally-driven seawater desalination: energy efficiency improvement by hybridization of the MED and AD cycles. Desalination 356:255–270 Ng KC, Shahzad MW, Son HS, Hamed QA (2017) An exergy approach to efficiency evaluation of desalination. Appl Phys Lett 110:184101 Pardo P, Deydier A, Anxionnaz-Minvielle Z, Rougé S, Cabassud M, Cognet P (2014) A review on high temperature thermochemical heat energy storage. Renew Sustain Energy Rev 32:591–610 Parmigiani L (2015) Water and energy in the GCC: securing scarce water in oil-rich countries. A report by The Institut français des relations internationales (Ifri). ISBN: 978-2-36567-442-3 Renewable Energy Market Analysis (2016) The GCC region. A report by International Renewable Energy Agency (IREA). ISBN: 978-92-95111-81-3 Shahzad MW (2013) The hybrid multi-effect desalination (MED) and the adsorption (AD) cycle for desalination. Doctoral Thesis, National University of Singapore Shahzad MW, Ng KC (2016) On the road to water sustainability in the Gulf, Nature Middle East. https://doi.org/10.1038/nmiddleeast.2016.50 Shahzad MW, Ng KC (2017) An improved multi-evaporator adsorption desalination cycle for GCC countries. Energy Technol 5:1663–1669 Shahzad MW, Thu K, Kim Y-D, Ng KC (2015) An experimental investigation on MEDAD hybrid desalination cycle. Appl Energy 148:273–281 Shahzad MW, Ng KC, Thu K (2016a) Future sustainable desalination using waste heat: kudos to thermodynamic synergy. Env Sci Water Res Technol 2:206–212 Shahzad MW, Thu K, Ng KC, WonGee C (2016b) Recent development in thermally activated desalination methods: achieving an energy efficiency less than 2.5 kWhelec/m3. Desalination Water Treat 57:7396–7405 Shahzad MW, Burhan M, Ang L, Ng KC (2017a) Energy-water-environment nexus underpinning future desalination sustainability. Desalination 413:52–64 Shahzad MW, Burhan M, Ghaffour N, Ng KC (2017b) A multi evaporator desalination system operated with thermocline energy for future sustainability. Desalination 435:268–277 Shahzad MW, Burhan M, Son HS, Oh SJ, Ng KC (2018) Desalination processes evaluation at common platform: a universal performance ratio (UPR) method. Appl Therm Eng 134:62–67 Shatat M, Riffat SB (2014) Water desalination technologies utilizing conventional and renewable energy sources. Int J Low-Carbon Technol 9:1–19 Shahzad MW, Ng KC, Thu, K, Saha BB, Chun WG (2014) Multi effect desalination and adsorption desalination (MEDAD): a hybrid desalination method. Appl Therm Eng 72:289– 297 Thu K, Kim Y-D, Shahzad MW, Saththasivam J, Ng KC (2015) Performance investigation of an advanced multi-effect adsorption desalination (MEAD) cycle. Appl Energy 159:469–477