Keith Lovegrove, Wes Stein-Concentrating solar power technology_

Keith Lovegrove, Wes Stein-Concentrating solar power technology_
Concentrating solar power technology
© Woodhead Publishing Limited, 2012
Related titles:
Materials for energy conversion devices
(ISBN 978-1-85573-932-1)
The term electroceramic is used to describe ceramic materials that have been
specially formulated with specific electrical, magnetic or optical properties.
Electroceramics are of increasing importance in many key technologies, including:
microelectronics, communications and energy conversion. This innovative book is
the first comprehensive survey on major new developments in electroceramics for
energy conversion devices. It presents current research from leading innovators in
the field.
Functional materials for sustainable energy applications
(ISBN 978-0-85709-059-1)
Functional materials are a class of advanced energy conversion materials of use in
photoelectric, thermoelectric, electrochemical, piezoelectric or electromagnetic
applications, such as photovoltaics (PV), hydrogen production and storage, fuel cell
systems, and demand-side energy management systems. Global demands for lower
cost, higher efficiency, mass production and, of course, sustainably sourced materials
have coupled with advances in nanotechnology to enable an increasingly important
role for functional materials in the sustainable energy mix. This book presents a
comprehensive review of the issues, science and development of functional materials in renewable and sustainable energy production and management applications.
Stand-alone and hybrid wind energy systems
(ISBN 978-1-84569-527-9)
Wind power generation is fast becoming one of the leading renewable energy
sources worldwide, with increasing penetration of stand-alone and hybrid wind
energy systems, particularly in distributed, isolated and community power networks.
Advanced energy storage and grid integration systems are required to provide
secure, reliable power supply to the end user. This book provides an extensive reference on the development of stand-alone and hybrid wind energy systems, as well as
energy storage and building-/grid-integration systems. Chapters cover the design,
construction, monitoring, control and optimisation of stand-alone and hybrid wind
energy technologies, and the continuing development of these systems.
Details of these books and a complete list of titles from Woodhead Publishing can
be obtained by:
•
•
•
visiting our web site at www.woodheadpublishing.com
contacting Customer Services (e-mail: [email protected]; fax: +44
(0) 1223 832819; tel.: +44 (0) 1223 499140 ext. 130; address: Woodhead Publishing
Limited, 80 High Street, Sawston, Cambridge CB22 3HJ, UK)
in North America, contacting our US office (e-mail: [email protected]
publishing.com; tel.: (215) 928 9112; address: Woodhead Publishing, 1518 Walnut
Street, Suite 1100, Philadelphia, PA 19102-3406, USA)
If you would like e-versions of our content, please visit our online platform: www.
woodheadpublishingonline.com. Please recommend it to your librarian so that
everyone in your institution can benefit from the wealth of content on the site.
© Woodhead Publishing Limited, 2012
Woodhead Publishing Series in Energy: Number 21
Concentrating solar
power technology
Principles, developments and applications
Edited by
Keith Lovegrove and Wes Stein
Oxford
Cambridge
Philadelphia
New Delhi
© Woodhead Publishing Limited, 2012
Published by Woodhead Publishing Limited,
80 High Street, Sawston, Cambridge CB22 3HJ, UK
www.woodheadpublishing.com
www.woodheadpublishingonline.com
Woodhead Publishing, 1518 Walnut Street, Suite 1100, Philadelphia,
PA 19102-3406, USA
Woodhead Publishing India Private Limited, G-2, Vardaan House, 7/28 Ansari
Road, Daryaganj, New Delhi – 110002, India
www.woodheadpublishingindia.com
First published 2012, Woodhead Publishing Limited
© Woodhead Publishing Limited, 2012; except Chapter 14 which was prepared by
US Government employees; it is therefore in the public domain and cannot be
copyrighted. Note: the publisher has made every effort to ensure that permission
for copyright material has been obtained by authors wishing to use such material.
The authors and the publisher will be glad to hear from any copyright holders it
has not been possible to contact.
The authors have asserted their moral rights.
This book contains information obtained from authentic and highly regarded
sources. Reprinted material is quoted with permission, and sources are indicated.
Reasonable efforts have been made to publish reliable data and information, but
the authors and the publisher cannot assume responsibility for the validity of all
materials. Neither the authors nor the publisher, nor anyone else associated with
this publication, shall be liable for any loss, damage or liability directly or
indirectly caused or alleged to be caused by this book.
Neither this book nor any part may be reproduced or transmitted in any form
or by any means, electronic or mechanical, including photocopying, microfilming
and recording, or by any information storage or retrieval system, without
permission in writing from Woodhead Publishing Limited.
The consent of Woodhead Publishing Limited does not extend to copying for
general distribution, for promotion, for creating new works, or for resale. Specific
permission must be obtained in writing from Woodhead Publishing Limited for
such copying.
Trademark notice: Product or corporate names may be trademarks or registered
trademarks, and are used only for identification and explanation, without intent to
infringe.
British Library Cataloguing in Publication Data
A catalogue record for this book is available from the British Library.
Library of Congress Control Number: 2012948529
ISBN 978-1-84569-769-3 (print)
ISBN 978-0-85709-617-3 (online)
ISSN 2044-9364 Woodhead Publishing Series in Energy (print)
ISSN 2044-9372 Woodhead Publishing Series in Energy (online)
The publisher’s policy is to use permanent paper from mills that operate a
sustainable forestry policy, and which has been manufactured from pulp which is
processed using acid-free and elemental chlorine-free practices. Furthermore, the
publisher ensures that the text paper and cover board used have met acceptable
environmental accreditation standards.
Typeset by Toppan Best-set Premedia Limited
Printed by Publishers’ Graphics LLC
© Woodhead Publishing Limited, 2012
Contents
Contributor contact details and author biographies
Woodhead Publishing Series in Energy
Foreword
Part I
Introduction
1
Introduction to concentrating solar power (CSP)
technology
K. LOVEGROVE, IT Power, Australia and W. STEIN,
CSIRO Energy Centre, Australia
Introduction
Approaches to concentrating solar power (CSP)
Future growth, cost and value
Organization of this book
References
1.1
1.2
1.3
1.4
1.5
2
2.1
2.2
2.3
2.4
2.5
2.6
2.7
2.8
2.9
2.10
2.11
2.12
2.13
Fundamental principles of concentrating solar
power (CSP) systems
K. LOVEGROVE, IT Power, Australia and J. PYE, Australian
National University, Australia
Introduction
Concentrating optics
Limits on concentration
Focal region flux distributions
Losses from receivers
Energy transport and storage
Power cycles for concentrating solar power (CSP) systems
Maximizing system efficiency
Predicting overall system performance
Economic analysis
Conclusion
Sources of further information and advice
References
xiii
xxiii
xxix
1
3
3
6
10
13
14
16
16
19
21
33
36
41
41
46
56
60
64
65
66
v
© Woodhead Publishing Limited, 2012
vi
Contents
3
Solar resources for concentrating solar power
(CSP) systems
R. MEYER, M. SCHLECHT and K. CHHATBAR,
Suntrace GmbH, Germany
Introduction
Solar radiation characteristics and assessment of solar
resources
Measuring solar irradiance
Deriving solar resources from satellite data
Annual cycle of direct normal irradiance (DNI)
Auxiliary meteorological parameters
Recommendations for solar resource assessment for
concentrating solar power (CSP) plants
Summary and future trends
References
3.1
3.2
3.3
3.4
3.5
3.6
3.7
3.8
3.9
4
4.1
4.2
4.3
4.4
4.5
4.6
4.7
5
5.1
5.2
5.3
5.4
5.5
5.6
Site selection and feasibility analysis for
concentrating solar power (CSP) systems
M. SCHLECHT and R. MEYER, Suntrace GmbH, Germany
Introduction
Overview of the process of site selection and feasibility
analysis
Main aspects considered during the pre-feasibility and
feasibility phases
Boundary conditions for a concentrating solar power (CSP)
project
Detailed analysis of a qualifying project location
Summary and future trends
References
Socio-economic and environmental assessment
of concentrating solar power (CSP) systems
N. CALDÉS and Y. LECHÓN, CIEMAT – Plataforma Solar de
Almería, Spain
Introduction
Environmental assessment of concentrating solar power
(CSP) systems
Socio-economic impacts of concentrating solar power (CSP)
systems
Future trends
Summary and conclusions
References
© Woodhead Publishing Limited, 2012
68
68
69
78
83
84
85
86
88
89
91
91
93
99
102
106
116
118
120
120
122
132
143
147
148
Contents
vii
Part II Technology approaches and potential
151
6
153
6.1
6.2
6.3
6.4
6.5
6.6
6.7
6.8
6.9
6.10
7
7.1
7.2
7.3
7.4
7.5
7.6
7.7
7.8
7.9
7.10
8
8.1
8.2
8.3
8.4
8.5
8.6
Linear Fresnel reflector (LFR) technology
D. R. MILLS, formerly Ausra Inc., Australia
Introduction
Historical background
Areva Solar (formerly Ausra, Solar Heat and Power)
Solar Power Group (formerly Solarmundo, Solel Europe)
Industrial Solar (formerly Mirroxx, PSE)
Novatec Solar (formerly Novatec-Biosol, Turmburg
Anlagenbau)
LFR receivers and thermal performance
Future trends
Conclusions
References
Parabolic-trough concentrating solar power (CSP)
systems
E. ZARZA MOYA, CIEMAT – Plataforma Solar de
Almería, Spain
Introduction
Commercially available parabolic-trough collectors (PTCs)
Existing parabolic-trough collector (PTC) solar thermal
power plants
Design of parabolic-trough concentrating solar power (CSP)
systems
Operation and maintenance (O&M) of parabolic-trough
systems
Thermal storage systems
Future trends
Conclusions
Sources of further information and advice
References and further reading
Central tower concentrating solar power (CSP)
systems
L. L. VANT-HULL, formerly University of Houston, USA
Introduction
History of central receivers
Activities since 2005
Design and optimization of central receiver systems
Heliostat factors
Receiver considerations
© Woodhead Publishing Limited, 2012
153
154
163
169
174
176
181
188
192
192
197
197
203
211
213
229
231
232
236
237
238
240
240
243
253
259
267
271
viii
Contents
8.7
8.8
8.9
8.10
8.11
8.12
Variants on the basic central receiver system
Field layout and land use
Future trends
Sources of further information and advice
Acknowledgements
References
9
Parabolic dish concentrating solar power (CSP)
systems
W. SCHIEL and T. KECK, schlaich bergermann und partner,
Germany
Introduction
Basic principles and historical development
Current initiatives
Energy conversion, power cycles and equipment
System performance
Optimization of manufacture
Future trends
Conclusion
Sources of further information and advice
References and further reading
9.1
9.2
9.3
9.4
9.5
9.6
9.7
9.8
9.9
9.10
10
10.1
10.2
10.3
10.4
10.5
10.6
10.7
10.8
11
11.1
11.2
Concentrating photovoltaic (CPV) systems and
applications
S. HORNE, SolFocus Inc., USA
Introduction
Fundamental characteristics of concentrating
photovoltaic (CPV) systems
Characteristics of high concentration photovoltaic (HCPV)
and low concentration photovoltaic (LCPV) devices and
their applications
Design of concentrating photovoltaic (CPV) systems
Examples of concentrating photovoltaic (CPV) systems
Future trends
Conclusions
References and further reading
Thermal energy storage systems for concentrating
solar power (CSP) plants
W.-D. STEINMANN, German Aerospace Center, Germany
Introduction: relevance of energy storage for concentrating
solar power (CSP)
Sensible energy storage
© Woodhead Publishing Limited, 2012
274
276
278
279
281
281
284
284
285
293
298
306
312
318
320
321
321
323
323
325
332
339
345
357
359
360
362
362
366
Contents
11.3
11.4
11.5
11.6
11.7
11.8
11.9
12
12.1
12.2
12.3
12.4
12.5
12.6
12.7
12.8
12.9
13
13.1
13.2
13.3
13.4
13.5
14
14.1
14.2
14.3
14.4
Latent heat storage concepts
Chemical energy storage
Selecting a storage system for a particular concentrating
solar power (CSP) plant
Future trends
Conclusion
Acknowledgement
References
ix
376
384
386
387
391
392
392
Hybridization of concentrating solar power (CSP)
with fossil fuel power plants
395
H. G. JIN and H. HONG, Chinese Academy of Sciences, China
Introduction
395
Solar hybridization approaches
396
Fossil boosting and backup of solar power plants
399
Solar-aided coal-fired power plants
402
Integrated solar combined cycle (ISCC) power plants
407
Advanced hybridization systems
412
Conclusions and future trends
418
Acknowledgements
419
References
419
Integrating a Fresnel solar boiler into an existing
coal-fired power plant: a case study
R. MILLAN, J. DE LALAING, E. BAUTISTA, M. ROJAS and
F GÖRLICH, Solar Power Group GmbH, Germany
Introduction
Description of options considered as variables selected
for the case study
Assessment of the solar add-on concept
Conclusions
References
The long-term market potential of concentrating
solar power (CSP) systems
S. J. SMITH, Pacific Northwest National Laboratory and
University of Maryland, USA
Introduction
Factors impacting the market penetration of concentrating
solar power (CSP)
Long-term concentrating solar power (CSP) market
potential
Summary and future trends
© Woodhead Publishing Limited, 2012
421
421
422
427
435
436
437
437
439
450
459
x
Contents
14.5
14.6
14.7
Sources of further information and advice
Acknowledgements
References
462
462
462
Part III Optimisation, improvements and applications
467
15
15.1
15.2
15.3
15.4
15.5
15.6
15.7
16
16.1
16.2
16.3
16.4
16.5
16.6
16.7
16.8
16.9
17
17.1
17.2
17.3
Absorber materials for solar thermal receivers in
concentrating solar power (CSP) systems
W. PLATZER and C. HILDEBRANDT, Fraunhofer Institute
for Solar Energy Systems, Germany
Introduction
Characterization of selective absorber surfaces
Types of selective absorbers
Degradation and lifetime
Examples of receivers for linearly concentrating collectors
Conclusion
References
Optimisation of concentrating solar power (CSP)
plant designs through integrated techno-economic
modelling
G. MORIN, Novatec Solar, Germany
Introduction
State-of-the-art in simulation and design of concentrating
solar power (CSP) plants
Multivariable optimisation of concentrating solar power
(CSP) plants
Case study definition: optimisation of a parabolic trough
power plant with molten salt storage
Case study results
Discussion of case study results
Conclusions and future trends
Acknowledgements
References
469
469
475
477
486
489
492
493
495
495
496
499
504
512
516
531
533
533
Heliostat size optimization for central receiver solar
power plants
536
J. B. BLACKMON, University of Alabama in Huntsville, USA
Introduction
536
Heliostat design issues and cost analysis
541
Category 1: costs constant per unit area irrespective of
heliostat size and number
546
© Woodhead Publishing Limited, 2012
17.4
17.5
17.6
17.7
17.8
17.9
18
18.1
18.2
18.3
18.4
18.5
18.6
19
19.1
19.2
19.3
19.4
19.5
19.6
19.7
20
20.1
20.2
20.3
Contents
xi
Category 2: size dependent costs
Category 3: fixed costs for each heliostat and other costs
Cost analysis as a function of area: the case of the 148 m2
Advanced Thermal Systems (ATS) glass/metal heliostat
Additional considerations in analysis of cost as a function
of area for the 148 m2 Advanced Thermal Systems (ATS)
glass/metal heliostat
Conclusion
References
548
555
Heat flux and temperature measurement
technologies for concentrating solar power (CSP)
J. BALLESTRÍN, CIEMAT – Plataforma Solar de Almería,
Spain and G. BURGESS and J. CUMPSTON, Australian
National University, Australia
Introduction
Heat flux measurement
Flux mapping system case studies
High temperature measurement
Conclusions
References
Concentrating solar technologies for industrial
process heat and cooling
A. HÄBERLE, PSE AG, Germany
Introduction
Technology overview
Components and system configuration
Case studies
Future trends and conclusion
Sources of further information and advice
References
Solar fuels and industrial solar chemistry
A. G. KONSTANDOPOULOS, Centre for Research and
Technology Hellas, Greece and Aristotle University,
Greece, C. PAGKOURA, Centre for Research and
Technology Hellas, Greece and University of West
Macedonia, Greece and S. LORENTZOU, Centre for
Research and Technology Hellas, Greece
Introduction
Solar chemistry
Hydrogen production using solar energy
© Woodhead Publishing Limited, 2012
557
565
574
575
577
577
578
587
593
598
598
602
602
603
606
612
616
618
618
620
620
623
626
xii
Contents
20.4
20.5
20.6
20.7
20.8
20.9
Solar-thermochemical reactor designs
Solar-derived fuels
Other applications of industrial solar chemistry
Conclusions
Acknowledgements
References
631
643
651
653
653
654
Index
662
© Woodhead Publishing Limited, 2012
Contributor contact details
and author biographies
(* = main contact)
Primary editor and Chapters 1* and 2*
Dr Keith Lovegrove (BSc 1984, PhD 1993) is currently Head – Solar
Thermal with the UK-based renewable energy consultancy group, IT Power.
He was previously Associate Professor and head of the solar thermal
group at the Australian National University where he led the team that
designed and built the 500 m2 generation II big dish solar concentrator. He
has served on the board of the ANZ Solar Energy Society as Chair, Vice
Chair and Treasurer. For many years he was Australia’s SolarPACES Task
II representative.
K. Lovegrove
IT Power
PO Box 6127 O’Connor
ACT 2602
Australia
E-mail: [email protected]
Editor and Chapter 1
Wes Stein is the Solar Energy Program Leader for CSIRO’s Division of
Energy Technology. He was responsible for establishing the National Solar
Energy Centre and has since grown a team of 30 engineers and scientists
and a strong portfolio of high temperature CSP research projects. He represents Australia on the IEA SolarPACES Executive Committee, and is a
member of the Australian Solar Institute Research Advisory Committee.
xiii
© Woodhead Publishing Limited, 2012
xiv
Contributor contact details
W. Stein
CSIRO Energy Centre
Steel River Eco Industrial Park
10 Murray Dwyer Close
Mayfield West
NSW 2304
Australia
E-mail: [email protected]
Chapter 2
John Pye is a researcher in the Australian National University Solar Thermal
Group and also lectures in the Department of Engineering.
J. Pye
Australian National University
Canberra
ACT 0200
Australia
E-mail: [email protected]
Chapter 3
Richard Meyer is co-founder and managing director of Germany-based
Suntrace. From 2006 to 2009, he headed the technical analysis and energy
yield teams of Epuron and SunTechnics. From 1996 to 2006, Richard worked
for DLR (German Aerospace Center), where he set up the satellite-based
services SOLEMI and DLR-ISIS for analyzing the potential for CSP. He
co-founded the IEA Task ‘Solar Resource Knowledge Management’, for
which he is the representative to the SolarPACES Executive Commitee. Dr
Richard Meyer holds a diploma in geophysics and a PhD in physics from
Munich University.
R. Meyer*, M. Schlecht and K. Chhatbar
Suntrace GmbH
Brandstwiete 46
20457 Hamburg
Germany
E-mail: [email protected]
Chapter 4
Martin Schlecht is co-founder and managing director of Germany-based
Suntrace, a highly specialized expert advisory firm in large scale solar. His
responsibilities include the assessment of CSP and PV project sites and
their feasibility. He has a Diploma (MSc) in mechanical engineering and
© Woodhead Publishing Limited, 2012
Contributor contact details
xv
more than 15 years’ work experience in the power industry, covering fossilfired, concentrating solar thermal and photovoltaic, including international
hands-on project development and project implementation.
M. Schlecht* and R. Meyer
Suntrace GmbH
Brandstwiete 46
20457 Hamburg
Germany
E-mail: [email protected]
Chapter 5
Natalia Caldés has a PhD in Agricultural and Natural Resources Economics from the Polytechnic University of Madrid and an MSc in Applied
Economics (University of Wisconsin-Madison). Her most relevant professional experience is in the field of development economics as well as energy
and enviromental economics. She joined the Spanish agency (CIEMAT) in
2004, where her work focuses on the socio-economic impact assessment of
energy technologies, evaluation of energy policies and energy modelling.
Yolanda Lechón has a PhD in Agricultural Engineering. She joined
CIEMAT in 1997. Her relevant experience involves life cycle assessment
and environmental externalities assessment of energy technologies and
energy modelling using techno-economic models.
N. Caldés and Y. Lechón*
Energy System Analysis Unit
Energy Department
CIEMAT – Plataforma Solar de Almería
Avda Complutense 22
28040 Madrid
Spain
E-mail: [email protected]
Chapter 6
David Mills has worked in non-imaging optics and solar concentrating
systems from 1976. At the University of Sydney, he ran the project that
created the double cermet selective absorber coating now used widely on
solar evacuated tubes and developed the CLFR concept. He was Cofounder, Chairman and CSO of both SHP P/L and Ausra Inc. (later Areva
Solar). He has been President of ISES (1997–99), first Chair of the International Solar Cities Initiative, and VESKI Entrepreneur in Residence for
the State of Victoria (2009).
© Woodhead Publishing Limited, 2012
xvi
Contributor contact details
D. R. Mills
Australia
Email: [email protected]
Chapter 7
Eduardo Zarza Moya is an Industrial Engineer with a PhD degree, born in
1958. At present he is the Head of the R&D Unit for Solar Concentrating
Systems at the Plataforma Solar de Almería in Spain. He has 27 years’
experience with solar concentrating systems, and has been the Director
of national and international R&D projects related to solar energy
and parabolic trough collectors. He is a member of the Scientific and Technical Committee of ESTELA (European Solar Thermal Electricity
Association).
E. Zarza Moya
CIEMAT – Plataforma Solar de Almería
Carretera de Tabernas a Senés, km 5
04200 Tabernas
Almería
Spain
E-mail: [email protected]
Chapter 8
Professor Lorin Vant-Hull has been involved in Solar Energy Projects since
1972. He retired as Professor Emeritus from the physics department of the
University of Houston in 2001, which he first joined in 1969. Dr Vant-Hull
was a Principal Investigator on the earliest US proposal to develop the
Solar Central Receiver project epitomized by the Solar One Pilot Plant (10
MWe at Barstow, California). He was program manager for eight years of
a Solar Thermal Advanced Research Center. Dr Vant-Hull has been an
Associate Editor for the Journal of Solar Energy for many years, as well as
a member of the Board of Directors of ASES and of ISES.
L. L. Vant-Hull
128 N Red Bud Trail
Elgin, TX 78621
USA
E-mail: [email protected]
Chapter 9
Wolfgang Schiel, Diplom Physicist, born in 1948 in Hamburg, has over 20
years’ experience in solar engineering, especially in design and construction
of several Dish/Stirling systems in Germany and other countries (Italy,
© Woodhead Publishing Limited, 2012
Contributor contact details
xvii
India, Spain, Turkey). After his degree at the University of Hamburg he
worked with the German Aerospace Research Establishment in Stuttgart.
In 1988 he joined schlaich bergermann und partner and became Managing
Director of sbp sonne gmbh in 2009.
Thomas Keck, Mechanical Engineer, born in 1959 in Stuttgart, joined schlaich bergermann und partner in 1988 and works as project manager for Dish/
Stirling projects.
W. Schiel* and T. Keck
schlaich bergermann und partner
Schwabstr. 43
70197 Stuttgart
Germany
E-mail: [email protected]
Chapter 10
Steve Horne is Co-Founder and Chief Technical Officer at SolFocus. He
began designing the concept of SolFocus’ CPV solar technology in 2005.
Before co-founding SolFocus, Steve was the Director of Engineering at
GuideTech, a leading semiconductor test equipment company, and had
previously spent six years running a technology consulting firm Tuross
Technology. He served as Vice President of Engineering at Ariel Electronics
and his early career experience includes commissioning two 500 MW steam
generated power plants in New South Wales, Australia.
S. Horne
SolFocus Inc.
510 Logue Avenue
Mountain View, CA 94043
USA
E-mail: [email protected]
Chapter 11
Dr Wolf-Dieter Steinmann has been working at the German Aerospace
Center (DLR) since 1994 and is project manager of the ‘CellFlux’ project
aiming at the development of an innovative thermal storage concept for
power plants. He was project manager of the European project DISTOR
and the national project PROSPER, which both deal with latent heat
storage for medium temperature applications. He completed his PhD thesis
on solar steam generators and has worked on the simulation and analysis
of the dynamics of thermodynamic systems.
© Woodhead Publishing Limited, 2012
xviii
Contributor contact details
W.-D. Steinmann
German Aerospace Center
Institute of Technical Thermodynamics
Pfaffenwaldring 38–40
70569 Stuttgart
Germany
E-mail: [email protected]
Chapter 12
Professor Hongguang Jin works in the field of solar thermal power technology and CO2 emission mitigation at the Institute of Engineering Thermophysics in Beijing. He established the mid-temperature solar thermochemical
process for integration of solar energy and fossil fuels, and originally proposed the Chemical Looping Combustion system with CO2 capture. He is
a past winner of the best paper award of ASME IGTI – international conference. He is a subject editor for the international journals Applied Energy
and Energy.
Dr Hui Hong is associate professor at the Institute of Engineering Thermophysics in Beijing. She works in the field of solar thermochemical
processing.
H. G. Jin* and H. Hong
Institute of Engineering Thermophysics
Chinese Academy of Sciences
Box 2706
Beijing 100190
China
E-mail: [email protected]; [email protected]
Chapter 13
Rosiel Millan has worked since 2009 at Solar Power Group GmbH as a
Process Engineer in charge of the design of solar thermal plants. She was
born in Mexico where she received her degree in Chemical Engineering.
She obtained her MSc in Renewable Energies from Carl von Ossietzky
Universität Oldenburg. She acquired her initial experience in thermal
energy storage systems for concentrating solar power plants while working
as research assistant at Fraunhofer-Institut für Solare Energiesysteme.
Count Jacques de Lalaing, founder of Solar Power Group and Managing
Director, is one of the pioneers of Fresnel solar power and established the
very first large-scale linear Fresnel pilot unit in the world in the 1990s. In
his former capacity as Chief Technology Officer at Solarmundo, Belgium,
© Woodhead Publishing Limited, 2012
Contributor contact details
xix
he raised awareness of the great potential of this new technology. In 2004,
he founded Solar Power Group.
R. Millan*, J. de Lalaing, E. Bautista, M. Rojas and F. Görlich
Solar Power Group GmbH
Daniel-Goldbach-Straße 17–19
40880 Ratingen
Germany
E-mail: [email protected]; [email protected]
Chapter 14
Steven J. Smith is a Senior Staff Scientist at the Joint Global Change
Research Institute, Pacific Northwest National Laboratory and University
of Maryland. His research focuses on energy systems, long-term socioeconomic scenarios and the interface between socioeconomic and climate
systems. Prior to joining PNNL in 1999, he worked at the National Center
for Atmospheric Research. Dr Smith was a lead author for the Intergovernmental Panel on Climate Change Special Report on Emissions Scenarios. He received his PhD in physics from UCLA.
S. J. Smith
Joint Global Change Research Institute
Pacific Northwest National Laboratory and University of Maryland
5825 University Research Court, Suite 3500
College Park, MD 20740
USA
E-mail: [email protected]
Chapter 15
Werner Platzer is division director ‘Solar Thermal and Optics’ at Fraunhofer ISE with more than 100 employees. Born in 1957, he graduated in
1982 in theoretical physics and acquired a PhD on solar gain and heat
transport in transparent insulation at the Albert-Ludwigs-University
Freiburg in 1988. He has been working in research and development of solar
thermal energy, facade technology and energy efficiency. He has authored
more than 150 articles and conference papers and lectures in solar thermal
energy at the University of Freiburg.
W. Platzer* and C. Hildebrandt
Fraunhofer Institute for Solar Energy Systems
Heidenhofstraße 2
79110 Freiburg
Germany
E-mail: [email protected]
© Woodhead Publishing Limited, 2012
xx
Contributor contact details
Chapter 16
Gabriel Morin has been working at Novatec Solar GmbH, Karlsruhe,
Germany, as a project manager in Research and Development since 2010.
From 2001 to 2010, he worked at the Fraunhofer Institute for Solar Energy
Systems (ISE) in the field of CSP, including as the coordinator of Solar
Thermal Power Plants. Gabriel Morin wrote his PhD thesis on technoeconomic design optimization of solar thermal power plants.
G. Morin
Novatec Solar
Herrenstraße 30
76133 Karlsruhe
Germany
E-mail: [email protected]
Chapter 17
James B. Blackmon (Aerospace Engineer, BS, 1961, Caltech; MS, 1967 and
PhD, 1972, UCLA) is currently Research Professor, Department of Mechanical and Aerospace Engineering, University of Alabama in Huntsville. He
was formerly Director, Product Development, McDonnell Douglas Corporation and Boeing Technical Fellow. He was the principal Investigator for
the DOE program for low cost heliostat development. His solar power
system experience began with a grant to McDonnell Douglas from NSF/
University of Houston for heliostat development in 1973. He has over 30
patents in space and terrestrial power, thermal management, and optical
and RF systems.
J. B. Blackmon
Department of Mechanical and Aerospace Engineering
University of Alabama in Huntsville
Huntsville, AL 35899
USA
E-mail: [email protected]
Chapter 18
Dr Jesus Ballestrín is currently researcher at Plataforma Solar de Almería
CIEMAT, Spain. He has more than 15 years’ experience of research and
development on solar concentrating technologies as central receivers, heliostats and solar furnaces. His research interests include: metrology of parameters related to concentrated solar radiation: high irradiances and high
superficial temperature.
Greg Burgess is the manager of the Solar Thermal Research Facility at the
Australian National University.
© Woodhead Publishing Limited, 2012
Contributor contact details
xxi
Jeff Cumpston is a PhD student in the Solar Thermal Group at the
Australian National University.
J. Ballestrín*
CIEMAT – Plataforma Solar de Almería
04200 Tabernas
Almería
Spain
E-mail: [email protected]
G. Burgess and J. Cumpston
Research School of Engineering
Building 31, North Rd
Australian National University
Canberra
ACT 0200
Australia
E-mail: [email protected]; [email protected]
Chapter 19
Dr Andreas Häberle is CEO of PSE AG, a spin-off company from the
Fraunhofer Institute for Solar Energy Systems ISE. Dr Häberle studied
Physics at the Technical University of Munich and then worked for seven
years as scientist and project manager at the Fraunhofer ISE where he
completed his PhD on concentrating solar thermal collectors before founding PSE in 1999. PSE AG specializes in solar test stands, solar consulting
and solar conference management.
A. Häberle
PSE AG
Emmy-Noether-Strasse 2
79110 Freiburg
Germany
E-mail: [email protected]
Chapter 20
Dr Athanasios G. Konstandopoulos is Founder and Director of APTL at
CPERI/CERTH (Greece) since 1996. He has served as Director of CPERI
(2006–2012) and since 2011 he is the Chairman of the Board and Managing
Director of CERTH. He is also Professor of New, Advanced & Clean Combustion Technologies at Aristotle University. He has a hybrid background
in Mechanical (Dipl. ME, AUTH, 1985; MSc ME Michigan Tech, 1987) and
Chemical Engineering (MSc, MPhil, PhD, Yale University, 1991) and
received the 2006 Descartes Laureate.
© Woodhead Publishing Limited, 2012
xxii
Contributor contact details
Chrysa Pagkoura is a Research Engineer at Aerosol and Particle Technology Laboratory of CPERI/CERTH and member of the HYDROSOL
research team.
Dr Souzana Lorentzou is an Affiliate Researcher at Aerosol and Particle
Technology Laboratory of CPERI/CERTH and member of the HYDROSOL research team.
A. G. Konstandopoulos*
Aerosol and Particle Technology Laboratory
Centre for Research and Technology Hellas
6th km Harillaou-Thermi Road
57001 Thermi-Thessaloniki
Greece
E-mail: [email protected]
and
Department of Chemical Engineering
Aristotle University
Thessaloniki
Greece
C. Pagkoura
Aerosol and Particle Technology Laboratory
Centre for Research and Technology Hellas
6th km Harillaou-Thermi Road
57001 Thermi-Thessaloniki
Greece
E-mail: [email protected]
and
Department of Mechanical Engineering
University of West Macedonia
Kozani 50100
Greece
S. Lorentzou
Aerosol and Particle Technology Laboratory
Centre for Research and Technology Hellas
6th km Harillaou-Thermi Road
57001 Thermi-Thessaloniki
Greece
E-mail: [email protected]
© Woodhead Publishing Limited, 2012
Woodhead Publishing Series in Energy
1
Generating power at high efficiency: Combined cycle technology for
sustainable energy production
Eric Jeffs
2
Advanced separation techniques for nuclear fuel reprocessing and
radioactive waste treatment
Edited by Kenneth L. Nash and Gregg J. Lumetta
3
Bioalcohol production: Biochemical conversion of lignocellulosic
biomass
Edited by K. W. Waldron
4
Understanding and mitigating ageing in nuclear power plants:
Materials and operational aspects of plant life management (PLiM)
Edited by Philip G. Tipping
5
Advanced power plant materials, design and technology
Edited by Dermot Roddy
6
Stand-alone and hybrid wind energy systems: Technology, energy
storage and applications
Edited by J. K. Kaldellis
7
Biodiesel science and technology: From soil to oil
Jan C. J. Bart, Natale Palmeri and Stefano Cavallaro
8
Developments and innovation in carbon dioxide (CO2) capture and
storage technology Volume 1: Carbon dioxide (CO2) capture,
transport and industrial applications
Edited by M. Mercedes Maroto-Valer
9
Geological repository systems for safe disposal of spent nuclear fuels
and radioactive waste
Edited by Joonhong Ahn and Michael J. Apted
10
Wind energy systems: Optimising design and construction for safe
and reliable operation
Edited by John D. Sørensen and Jens N. Sørensen
xxiii
© Woodhead Publishing Limited, 2012
xxiv
Woodhead Publishing Series in Energy
11
Solid oxide fuel cell technology: Principles, performance and
operations
Kevin Huang and John Bannister Goodenough
12
Handbook of advanced radioactive waste conditioning technologies
Edited by Michael I. Ojovan
13
Membranes for clean and renewable power applications
Edited by Annarosa Gugliuzza and Angelo Basile
14
Materials for energy efficiency and thermal comfort in buildings
Edited by Matthew R. Hall
15
Handbook of biofuels production: Processes and technologies
Edited by Rafael Luque, Juan Campelo and James Clark
16
Developments and innovation in carbon dioxide (CO2) capture and
storage technology Volume 2: Carbon dioxide (CO2) storage and
utilisation
Edited by M. Mercedes Maroto-Valer
17
Oxy-fuel combustion for power generation and carbon dioxide (CO2)
capture
Edited by Ligang Zheng
18
Small and micro combined heat and power (CHP) systems:
Advanced design, performance, materials and applications
Edited by Robert Beith
19
Advances in clean hydrocarbon fuel processing: Science and
technology
Edited by M. Rashid Khan
20
Modern gas turbine systems: High efficiency, low emission, fuel
flexible power generation
Edited by Peter Jansohn
21
Concentrating solar power technology: Principles, developments and
applications
Edited by Keith Lovegrove and Wes Stein
22
Nuclear corrosion science and engineering
Edited by Damien Féron
23
Power plant life management and performance improvement
Edited by John E. Oakey
24
Direct-drive renewable energy systems
Edited by Markus Mueller and Henk Polinder
© Woodhead Publishing Limited, 2012
Woodhead Publishing Series in Energy
xxv
25
Advanced membrane science and technology for sustainable energy
and environmental applications
Edited by Angelo Basile and Suzana Pereira Nunes
26
Irradiation embrittlement of reactor pressure vessels (RPVs) in
nuclear power plants
Edited by Naoki Soneda
27
High temperature superconductors (HTS) for energy applications
Edited by Ziad Melhem
28
Infrastructure and methodologies for the justification of nuclear
power programmes
Edited by Agustín Alonso
29
Waste to energy (WtE) conversion technology
Edited by Marco Castaldi
30
Polymer electrolyte membrane and direct methanol fuel cell
technology Volume 1: Fundamentals and performance of low
temperature fuel cells
Edited by Christoph Hartnig and Christina Roth
31
Polymer electrolyte membrane and direct methanol fuel cell
technology Volume 2: In situ characterization techniques for low
temperature fuel cells
Edited by Christoph Hartnig and Christina Roth
32
Combined cycle systems for near-zero emission power generation
Edited by Ashok D. Rao
33
Modern earth buildings: Materials, engineering, construction and
applications
Edited by Matthew R. Hall, Rick Lindsay and Meror Krayenhoff
34
Metropolitan sustainability: Understanding and improving the urban
environment
Edited by Frank Zeman
35
Functional materials for sustainable energy applications
Edited by John Kilner, Stephen Skinner, Stuart Irvine and
Peter Edwards
36
Nuclear decommissioning: Planning, execution and international
experience
Edited by Michele Laraia
37
Nuclear fuel cycle science and engineering
Edited by Ian Crossland
© Woodhead Publishing Limited, 2012
xxvi
Woodhead Publishing Series in Energy
38
Electricity transmission, distribution and storage systems
Edited by Ziad Melhem
39
Advances in biodiesel production: Processes and technologies
Edited by Rafael Luque and Juan A. Melero
40
Biomass combustion science, technology and engineering
Edited by Lasse Rosendahl
41
Ultra-supercritical coal power plant: Materials, technologies and
optimisation
Edited by Dongke Zhang
42
Radionuclide behaviour in the natural environment: Science,
implications and lessons for the nuclear industry
Edited by Christophe Poinssot and Horst Geckeis
43
Calcium and chemical looping technology for power generation and
carbon dioxide (CO2) capture: Solid oxygen- and CO2-carriers
P. Fennell and E. J. Anthony
44
Materials’ ageing and degradation in light water reactors:
Mechanisms, modelling and mitigation
Edited by K. L. Murty
45
Structural alloys for power plants: Operational challenges and hightemperature materials
Edited by Amir Shirzadi, Rob Wallach and Susan Jackson
46
Biolubricants: Science and technology
Jan C. J. Bart, Emanuele Gucciardi and Stefano Cavallaro
47
Wind turbine blade design and materials: Improving reliability, cost
and performance
Edited by Povl Brøndsted and Rogier Nijssen
48
Radioactive waste management and contaminated site clean-up:
Processes, technologies and international experience
Edited by William E. Lee, Michael I. Ojovan, Carol M. Jantzen
49
Probabilistic safety assessment for optimum nuclear power plant life
management (PLiM)
Gennadij V. Arkadov, Alexander F. Getman and Andrei N. Rodionov
50
Coal utilization in industry
Edited by D. G. Osborne
51
Coal power plant materials and life assessment: Developments and
applications
Edited by Ahmed Shibli
© Woodhead Publishing Limited, 2012
Woodhead Publishing Series in Energy
xxvii
52
The biogas handbook: Science, production and applications
Edited by Arthur Wellinger and David Baxter
53
Advances in biorefineries: Biomass and waste supply chain
exploitation
Edited by K. W. Waldron
54
Geoscience of carbon dioxide (CO2) storage
Edited by Jon Gluyas and Simon Mathias
55
Handbook of membrane reactors Volume 1: Fundamental materials
science, design and optimisation
Edited by Angelo Basile
56
Handbook of membrane reactors Volume 2: Industrial applications
and economics
Edited by Angelo Basile
57
Alternative fuels and advanced vehicle technologies: Towards zero
carbon transportation
Edited by Richard Folkson
58
Handbook of microalgal bioprocess engineering
Christopher Lan and Bei Wang
59
Fluidized-bed technologies for near-zero emission combustion and
gasification
Edited by Fabrizio Scala
60
Managing nuclear projects: A comprehensive management resource
Edited by Jas Devgun
61
Handbook of process integration: Energy, water, waste and emissions
management in processing and power industries
Edited by Jiří Klemeš
© Woodhead Publishing Limited, 2012
Foreword
During this century the human race will have to address the challenge of
deeply transforming the world energy system to make it much more sustainable and environmentally friendly than the one we currently have. To
achieve this, it will have to substantially increase the market penetration of
all types of renewable energy technologies, and especially of solar technologies, since these technologies will be called upon to be the main pillars of
the new world energy system, because of the vast quantities and the high
quality of the solar energy reaching the Earth at every instant.
The shift towards a much greener world energy system requires an
extraordinary mobilization of technological and economic resources. The
good news is that this mobilization is starting to happen. According to the
US Department of Energy, in 2011, for the first time in history, worldwide
investment in renewable electricity generation capacity exceeded the
worldwide investment in conventional systems.
To enable the required large-scale development and deployment of
renewable energy systems worldwide, it is essential to ensure that the
renewable energy industry has access to affordable finance and to the necessary renewable energy expertise and know-how.
This book represents an important contribution to disseminate the
knowledge and expertise that its authors have in the field of concentrating
solar power (CSP). The diversity of countries, institutions and fields of
expertise represented by the contributors to this book, and the quality of
their contributions also constitute an example in itself of the rapid but solid
expansion that the CSP international community has undergone over recent
decades.
In addition to congratulating the editors and the authors for delivering
this excellent book, I would like to end this foreword by pointing out the
fact that many of the contributors to this book and their institutions are
active participants in the activities of SolarPACES, the Implementing
Agreement of the International Energy Agency for ‘Solar Power and
xxix
© Woodhead Publishing Limited, 2012
xxx
Foreword
Chemical Energy Systems’. This is not by chance; the rapid expansion that
the CSP industry is experiencing worldwide since 2003 owes much to the
unfaltering work of SolarPACES over the last 30 years.
Manuel J. Blanco, PhD Dr Ing.
Chair, SolarPACES Executive Committee
© Woodhead Publishing Limited, 2012
1
Introduction to concentrating solar power
(CSP) technology
K. L O V E G R O V E, IT Power, Australia and
W. S T E I N, CSIRO Energy Centre, Australia
Abstract: This introductory chapter begins by defining ‘concentrating
solar power’ (CSP) and outlining the role of the book. It then introduces
some of the historical background to the development of CSP systems
and the present day context of a period of industry growth amid major
changes to the world’s energy systems. It describes the key approaches
of parabolic trough, central receiver, linear Fresnel, Fresnel lens and
paraboloidal dish concentrator systems. The prospects for continued
deployment growth and parallel cost reductions are discussed. Finally
the organization of the overall book is outlined.
Key words: concentrating solar power, concentrating photovoltaics, dish,
trough, tower, Fresnel lens, linear Fresnel reflector, history, approaches to
concentration, cost reduction, growth in deployment.
1.1
Introduction
Concentrating solar power (CSP) systems use combinations of mirrors or
lenses to concentrate direct beam solar radiation to produce forms of useful
energy such as heat, electricity or fuels by various downstream technologies.
The term ‘concentrating solar power’ is often used synonymously with
‘concentrating solar thermal power’. In this book the term is used in a more
general sense to include both concentrating solar thermal (CST) and concentrating photovoltaic (CPV) energy conversion.
Whilst the primary commercial attention today and the emphasis in this
book is on systems designed for generation of electric power, there are
individual chapters that review the important market segment of process
heat and also the concept of solar fuels production, which the editors
suggest is likely to see a rapid rise in interest in the near future.
This book seeks to address multiple audiences, and chapters can be read
selectively according to need.
•
•
A reader with a background in science or engineering should find a
resource that introduces all the key principles and the state of the art
of the CSP field.
Many of the chapters contain detailed review and presentation on
various key aspects that should provide value to those experts already
3
© Woodhead Publishing Limited, 2012
4
•
Concentrating solar power technology
working in the field and, given the pace of technological change, suggested resources for remaining up to date.
At the same time, the book should provide value to readers without a
technical background. Care has been taken to provide overviews and
introductions of all key concepts in a manner targeted at the nontechnical audience such as policy makers, for example.
This book seeks to provide comprehensive, complete and up-to-date coverage of the CSP field. A previous well-respected coverage of this nature was
provided by Winter et al. (1991). There are a number of past and recent
books that address broader solar energy topics and others with more technical coverage of specific issues, which are referenced in various chapters
where relevant.
1.1.1
History and context
Global investments in clean energy generation are continuing to increase
with global energy producers (and users) now experiencing strong signals to
develop a clean energy future. Over the last three decades, the world wind
industry has grown at an average rate of approximately 30% per year to
reach a total installed capacity of 239 GW by the end of 2011. This represents
nearly 3% of total world electricity annual generation (WWEA, 2012) and
wind capacity is now being installed at a faster annual rate than nuclear.
Over a shorter period, the solar photovoltaic (PV) industry has grown
with comparable or higher rates of growth but from a lower base and in
2011 had a worldwide installed capacity of approximately 69 GW (EPIA,
2012). CSP technology saw a first surge of commercial development between
1984 and 1995, but then no further commercial deployment until 2005,
although in that time considerable research, development and demonstration took place. Since then, commercial CSP deployment has recommenced
and gained considerable momentum. Total installed capacity is, however, an
order of magnitude smaller than PV, given that commercialization of the
technology is a decade or so behind.
The concept of concentrating solar energy has been a technology of interest throughout history. For example:
•
•
•
•
Archimedes described the idea of mirrored panels to concentrate the
sun in around 200 BC;
The Greek mathematician Diocles described the optical properties of a
parabolic trough in the second century BC;
The development of heliostat designs was described by Comte de Buffon
in 1746;
Augustin Mouchot demonstrated a dish driven steam engine system at
the 1878 universal exhibition in Paris.
© Woodhead Publishing Limited, 2012
Introduction to concentrating solar power (CSP) technology
5
A more contemporary historical landmark was Frank Schuman’s successful
parabolic trough driven pumping system built in Egypt in 1913. Experiments and prototypes were developed all through the twentieth century.
The real birth of CSP as an industry came in California in the 1980s. Favourable government policy settings lead to the construction of nine separate
parabolic trough based ‘Solar Electric Generating Systems’ (SEGS), totalling 354 MWe of installed capacity. These were based around steam turbines
for power generation, and used oil as the heat transfer fluid within the
trough receivers.
These plants, with more than 2,000,000 m2 of mirror area, continue to
operate under utility ownership after more than 20 years and have established the technology as commercially proven. The tenth plant was in the
early stages of construction when the effect of lower oil prices and changes
in government policy led to a loss of investment and subsequent demise of
the company driving the development (LUZ). However, the technology was
now on the map, and over that 1984–95 period, with just 354 MW deployed,
the capital cost was successfully halved.
The lead role in renewable energy development was grasped around that
time by countries in north-western Europe, led by Denmark and then
Germany. The emphasis was on pursuing wind power given the favourable
wind and less favourable solar resources in those countries. Though wind
turbines today are of the order of 3–5 MW per unit, at that time they were
in the small hundreds of kW, and even though the specific capital cost was
similar to or higher than CSP, the smaller modules provided a much easier
investment path. Led by government incentives, PVs have moved from high
cost space/satellite and small remote off-grid applications to residential
applications and more recently large multi-MW installations. The renewable energy agenda has spread around the globe and overall market demand
for renewable electricity continues to grow exponentially, though the ‘new’
renewables such as wind and PV still account for only a few percent of the
world’s electricity demand.
A past and continuing challenge for CSP is its dependence on the economies of scale afforded by large steam turbines, leading to large levels of risk
capital per project for a relatively new technology. However, now that the
size of new renewable projects has grown, there is more appetite for making
the necessary investments.
Concern over human induced climate change has emerged to dominate
the political agenda around energy supply. There has been a resurgence
of CSP development since 2005, led partly by the recognition that it is a
technology which could make large greenhouse gas emission cuts quickly,
and offer the significant benefit of distributable solar power through
integrated thermal storage. This growth has been led predominantly by
Spain through specific and targeted feed-in tariff incentives that have
© Woodhead Publishing Limited, 2012
6
Concentrating solar power technology
proven highly successful for the technology. Approximately 2,400 MW is
approved for operation by 2014 with half of that already operating. The
sun belt of the south-west USA has also been targeted for CSP through
tax credits and loan guarantees with approximately 1.8 GW expected to
be in operation by the end of 2013. Importantly, the majority of new installations now incorporate thermal storage, usually of the order of 6 hours
or so.
Other countries with CSP projects announced or under construction
include North Africa (Algeria, Morocco) and the Middle East (Egypt,
Israel), China, India, Australia, South Africa, Portugal, Italy, Greece, Malta
and Cyprus. In 2010, India took a major initiative with the establishment of
the Jawaharlal Nehru National Solar Mission, with a target of 20 GWe of
combined PV and CSP capacity to be installed by 2022. China has a target
of 1 GW of CSP by 2015. This activity has combined to give a rate of growth
from 2005 to 2012 of approximately 40% per year. This is similar to the rate
of growth for wind power during its first decade of modern commercial
deployment, which began in approximately 1990, and faster than that for
PVs when it began to accelerate commercial deployment in about 1992.
Whilst the industry is still in its early stages and vulnerable to sudden policy
changes in key countries, continued strong growth in global installed capacity is predicted.
Due to the 15-year hiatus in commercial CSP deployments, installed PV
capacity grew to be some ten times greater than CSP, and as a result PV
has seen significant cost reduction over recent years, whilst CSP is at an
early stage of its cost reduction path. In 2012, PV is lower cost than CSP
for non-dispatchable electricity production under most applications. Under
these circumstances, greater attention is turning to CSP’s potential benefits
of built-in thermal energy storage and dispatchability, as well as other nonelectrical applications such as fuels.
Whilst the issue of climate change is dominating the future energy agenda,
the idea that demand for oil may have now passed the level of supply from
conventional sources is well accepted and, despite large levels of fluctuation,
the overall trend is to increasing prices. This could prove to be a very major
driver for technology change both increasing demand for solar electricity
and encouraging developments such as solar fuels.
1.2
Approaches to concentrating solar power (CSP)
CSP systems capture the direct beam component of solar radiation. Unlike
flat plate photovoltaics (PV), they are not able to use radiation that has
been diffused by clouds or dust or other factors. This makes them best suited
to areas with a high percentage of clear sky days, in locations that do not
have smog or dust.
© Woodhead Publishing Limited, 2012
Introduction to concentrating solar power (CSP) technology
7
The configurations that are currently used commercially in order of
deployment level are:
•
•
•
•
•
parabolic trough
central receiver tower
linear Fresnel
Fresnel lenses (for CPV)
paraboloidal dishes.
Each technology boasts particular advantages and in some cases particular
market segments. Project and technology developers are actively pursuing
all types of CSP technologies. In addition to these concepts that are applied
commercially, a solar furnace arrangement is widely used as a tool for
research projects. A solar furnace typically consists of a paraboloidal dish
mounted in a fixed orientation in a laboratory building, with one or more
external heliostats directing solar radiation to it at a fixed angle.
1.2.1
Parabolic trough
Parabolic trough-shaped mirrors produce a linear focus on a receiver tube
along the parabola’s focal line as illustrated in Fig. 1.1. The complete assembly of mirrors plus receiver is mounted on a frame that tracks the daily
movement of the sun on one axis. Relative seasonal movements of the sun
in the other axis result in lateral movements of the line focus, which remains
on the receiver but can have some spill at the row ends.
1.1 Parabolic trough collector: tracks the sun on one axis (background
picture, Nevada Solar 1 plant, R. Dunn).
© Woodhead Publishing Limited, 2012
8
Concentrating solar power technology
1.2 Central receiver tower plant: multiple heliostats move on two axes
to focus the sun to a fixed tower mounted receiver (background
picture, Gemasolar plant, owned by Torresol Energy, © Torresol
Energy).
Trough systems using thermal energy collection via evacuated tube
receivers are currently the most widely deployed CSP technology. In this
configuration, an oil heat transfer fluid is usually used to collect the heat
from the receiver tubes and transport it to a central power block. Chapter
7 examines trough systems in detail.
1.2.2
Central receiver tower
A central receiver tower system involves an array of heliostats (large
mirrors with two axis tracking) that concentrate the sunlight onto a fixed
receiver mounted at the top of a tower, as illustrated in Fig. 1.2. This allows
sophisticated high efficiency energy conversion at a single large receiver
point. Higher concentration ratios are achieved compared to linear focusing
systems and this allows thermal receivers to operate at higher temperatures
with reduced losses. A range of system and heliostat sizes have been demonstrated. Chapter 8 examines tower systems in detail.
1.2.3
Linear Fresnel reflectors
Linear Fresnel reflector (LFR) systems produce a linear focus on a downward facing fixed receiver mounted on a series of small towers as shown in
Fig. 1.3. Long rows of flat or slightly curved mirrors move independently
on one axis to reflect the sun’s rays onto the stationary receiver. For thermal
© Woodhead Publishing Limited, 2012
Introduction to concentrating solar power (CSP) technology
9
1.3 Linear Fresnel reflector: multiple mirrors move on one axis to
focus the sun to a fixed linear receiver (background picture,
Kimberlina LFR plant, Bakersfield California, image courtesy of AREVA
Solar).
systems, the fixed receiver not only avoids the need for rotary joints for the
heat transfer fluid, but can also help to reduce convection losses from a
thermal receiver because it has a permanently down-facing cavity.
The proponents of the LFR approach argue that its simple design with
near flat mirrors and less supporting structure, which is closer to the ground,
outweighs the lower overall optical and (for CST) thermal efficiency. To
increase optical and ground-use efficiency, compact linear Fresnel reflectors
(CLFRs) use multiple receivers for each set of mirrors so that adjacent
mirrors have different inclinations in order to target different receivers. This
allows higher packing density of mirrors which increases optical efficiency
and minimizes land use. Chapter 6 examines linear Fresnel systems in detail.
1.2.4
Fresnel lens
A conventional lens is expensive and impractical to manufacture on a large
scale. The Fresnel lens overcomes these difficulties and has been employed
extensively for CPV systems. A Fresnel lens is made as a series of concentric
small steps, each having a surface shape matching that which would be
found on a standard lens but with all the steps kept within a small thickness.
A plastic material is usually used and arrays of multiple lens units are typically mounted on a heliostat structure as shown in Fig. 1.4. This is also a
© Woodhead Publishing Limited, 2012
10
Concentrating solar power technology
Fresnel
lens
Target
(single
cell)
1.4 Fresnel lens-based CPV: multiple small units on a heliostat
(background picture, River Mountains, USA, Amonix).
point focus approach requiring accurate sun tracking in two axes. Chapter
10 examines various CPV systems in detail.
1.2.5
Parabolic dishes
Dish systems, like troughs, exploit the geometric properties of a parabola,
but as a three-dimensional paraboloid as shown in Fig. 1.5. The reflected
direct beam radiation is concentrated to a point focus receiver and in CST
systems can heat this to operating temperatures of over 1,000ºC, similar to
tower systems.
Dish systems offer the highest potential solar conversion efficiencies of
all the CSP technologies, because they always present their full aperture
directly towards the sun and avoid the ‘cosine loss effect’ that the other
approaches experience. They are, however, the least commercially mature.
Dishes up to 24 m diameter have been demonstrated.
As well as thermal conversion, CPV conversion on dishes is well established, it is also applied with ‘micro dishes’ with diameters of just several
centimetres. Chapter 9 examines dish systems in detail.
1.3
Future growth, cost and value
CSP systems produce renewable electricity that ultimately must compete
with other forms of electricity generation in the marketplace. Thus the cost
© Woodhead Publishing Limited, 2012
Introduction to concentrating solar power (CSP) technology
11
1.5 Paraboloidal dish concentrator: tracks the sun in two axes
(background picture, Australian National University, 500 m2 dish).
of CSP energy is the main preoccupation of the technology developers and
research and development practitioners within the CSP community. With
no fuel costs, the cost of CSP energy is dominated by the amortization of
the high initial capital cost investment over the life of the plant.
CSP is a proven technology that is at an early stage of its cost reduction
curve. A period of rapid growth in installed capacity, together with a rapid
decay in cost of energy produced is confidently predicted by the industry.
The trend of cost reduction as installed capacity increases is logically
linked to:
•
•
•
technical improvements, as lessons are learned from installed plants and
parallel R&D efforts identify performance improvements,
scaling to larger installed plant size, which allows for more efficient and
more cost-effective large turbines and other components to be used, and
volume production that allows fixed costs of investments in production
efficiency to be spread over larger production runs.
Empirically these practical effects lead to a commonly observed trend for
a new technology of a reduction in cost of an approximately fixed fraction
for every doubling of deployed capacity.
An analysis of various comprehensive studies investigating feasible cost
reduction paths for CSP was carried out in a study for the Global Environment Facility for the World Bank in 2006 (World Bank 2006). One comprehensive scenario predicted a pathway to install 5 GW by 2015.
© Woodhead Publishing Limited, 2012
12
Concentrating solar power technology
Capacity (GWe)
5
4
Actual
19%/yr
25%/yr
40%/yr
3
2
1
0
1980
1990
2000
2010
2020
2030
Year
1.6 Global installed capacity of CSP plants, both actual and possible
future compound growth rates.
Capacity (GWe)
100
Actual
19%/yr
25%/yr
40%/yr
Wind
PV
80
60
40
20
0
1980
1990
2000
2010
2020
2030
Year
1.7 Global installed capacity of CSP plants, both actual and possible
future compound growth rates together with historical data for wind
and PV deployment.
A recent roadmap published by the International Energy Agency (IEA)
for CSP technology presents a highly credible summary of the global situation and way forward (IEA 2010). Cost of energy reductions to around
25% of 2010 values are predicted by 2050. AT Kearney (2010) was commissioned by European and Spanish CST industry associations to produce a
study of CSP energy cost reduction projections. A range of key areas for
reducing cost of manufacture and increasing annual output are identified,
these measures together are suggested to result in an overall reduction of
cost of energy in 2025 relative to 2012 of 40–50%. Over the same time
period, they suggest global installed capacity could reach between 60 and
100 GW depending on policy measures in place. Figure 1.6 illustrates the
history of installed capacity to 2011 together with extrapolations based on
compound growth rates of the 19% per year average since 1984 and the
40% per year average since 2005.
Figure 1.7 shows the same data on an expanded vertical axis, together
with actual historical data for installed capacity of wind and PV systems.
The historical high compound growth rates for these technologies can be
seen together with the approximately one decade lag between PV growth
© Woodhead Publishing Limited, 2012
Introduction to concentrating solar power (CSP) technology
13
Relative LCOE
1.0
20%/a + PR = 0.8
20%/a + PR = 0.85
20%/a + PR = 0.9
30%/a + PR = 0.8
30%/a + PR = 0.85
30%/a + PR = 0.9
0.8
0.6
0.4
0.2
0.0
2010
2015
2020
Year
2025
2030
1.8 Possible relative levelized cost of energy (LCOE) reductions over
time under different growth rates and progress ratios.
and that of wind. CSP is seen to be entering a similar growth phase with a
further approximately one decade lag.
Other studies support the conclusions on cost reduction potential.
Kutscher et al. (2010) identifies in detail a range of specific ‘bottom-up’
measures that are estimated to deliver a 40% cost of energy reduction for
line focus systems by 2017. Kolb et al. (2010) identifies measures that will
deliver 50% cost reductions for tower systems by 2020.
Available evidence points to a cost reduction of 10–15% for every doubling in global capacity (a progress ratio of 0.9–0.85). Figure 1.8 plots the
progression over time of relative costs (either cost of energy or capital
costs1) under either 20% p.a. or 30% p.a. growth rates, and for cost progress
ratios of 0.8, 0.85 and 0.9.
As variable renewables like wind and PV vie for a larger proportion of
energy supply, the ability to provide dispatchable power will become more
important. CSP has the advantages that incorporation of thermal energy
storage is cost effective, improves system performance and has very little
effect on the overall cost of energy. Energy storage is examined in detail in
Chapter 11. Some recent studies have begun to evaluate the extra value
that can be offered by the energy storage abilities of CSP systems (e.g.
Sioshansi and Denholm, 2010 and Madaeni et al., 2011), and it can be 30%
or more valuable than average market prices. Thus CSP can look forward
to a growing recognition of the value of its energy in parallel with future
cost reductions.
1.4
Organization of this book
The book is organized into three parts.
1
Note that cost of energy is strongly dependent on capital cost, but also depends on operating
and maintenance costs and financing costs. For a first approximation, cost of energy and capital
cost are assumed to reduce over time according to the same progress ratio.
© Woodhead Publishing Limited, 2012
14
Concentrating solar power technology
Part I contains fundamental introductory material of which this introductory chapter forms the first part. This is followed by Chapter 2 which overviews the fundamental principles behind CSP technologies. It is quite a
technical chapter that can be skipped by those seeking to read directly
about specific technology. Understanding solar resources issues, siting and
feasibility studies and the techno-economic assessment of CSP systems are
the subject of the other chapters in Part I.
Part II, on technology approaches and potential, contains specific chapters that review the principles, historical development and state of the art
of the trough, tower, linear Fresnel and dish approaches. These are followed
by further chapters on energy storage, hybridization, CPV systems and
finally the economic outlook.
Part III, on optimization, improvements and applications, comprises
chapters that provide in-depth coverage of a range of key issues around
maximizing performance through technology and design optimization. Key
applications to process heat and solar fuels are also presented as a complement to the overall emphasis on power generation. Solar fuels derived from
concentrating solar systems are presented as the last chapter of the book.
This reflects a belief on the part of the editors that whilst solar fuels is currently an activity still very much in the R&D sphere, it could well become
the biggest future market for solar concentrating systems, given future
projections of demand outstripping supply for oil.
1.5
References
AT Kearney (2010), Solar Thermal Electricity 2025, Report for ESTELLA by
A.T. Kearney GmbH, Kaistraße 16A, 40221 Duesseldorf, Germany.
IEA (2010), Technology Roadmap – Concentrating Solar Power, OECD International Energy Agency, Publications Service, OECD 2, rue André-Pascal, 75775
Paris cedex 16, France.
EPIA (2012), Global market outlook for photovoltaics until 2016, European Photovoltaic Industry Association, Rue d’Arlon 63-67 (1040) Brussels – Belgium, www.
epia.org.
Kolb G, Ho C, Mancini T and Gary J (2011), Power Tower Technology Roadmap
and Cost Reduction Plan, SANDIA REPORT SAND2011-2419, April 2011, Prepared by Sandia National Laboratories Albuquerque, NM 87185 and Livermore,
CA 94550, http://prod.sandia.gov/techlib/access-control.cgi/2011/112419.pdf.
Kutscher C, Mehos M, Turchi C, Glatzmaier G and Moss T (2010), Line-Focus Solar
Power Plant Cost Reduction Plan, NREL/TP-5500-48175, December, http://www.
nrel.gov/docs/fy11osti/48175.pdf.
Madaeni S Sioshansi R and Denholm P (2011), Capacity Value of Concentrating
Solar Power Plants, National Renewable Energy Laboratory Technical Report
NREL/TP-6A20-51253.
Sioshansi R and Denholm P (2010), The Value of Concentrating Solar Power and
Thermal Energy Storage, Technical Report NREL-TP-6A2-45833, February.
© Woodhead Publishing Limited, 2012
Introduction to concentrating solar power (CSP) technology
15
Winter C-J, Sizmann R L and Vant-Hull L L (1991), Solar Thermal Power Plants,
New York, Springer Verlag.
World Bank (2006), Assessment of the World Bank/GEF Strategy for the Market
Development of Concentrating Solar Thermal Power, Report Prepared for the
World Bank.
WWEA (2012), Wind energy around the world, World Wind Energy Association,
quarterly report, Editor-in-Chief: Stefan Gsänger, Issue 1, March.
© Woodhead Publishing Limited, 2012
2
Fundamental principles of concentrating solar
power (CSP) systems
K. L O V E G R O V E, IT Power, Australia and
J. P Y E, Australian National University, Australia
Abstract: This chapter provides an overview of the fundamental
principles of CSP systems. It begins with the optical processes and the
ultimate limits on the extent to which solar radiation can be
concentrated. Practical factors that reduce achievable concentration
levels further are discussed. Mechanisms of thermal energy loss from
receivers are covered. Available power cycles for electricity generation
are reviewed. The second law of thermodynamics is introduced to lead
into a consideration of optimization of overall system efficiency via
variation of operating temperature and receiver aperture size.
Performance modelling of complete systems is introduced and finally the
analysis of levelized cost of energy is covered, as a metric for comparing
systems, and as a tool to thermo-economic optimization in design.
Key words: maximum concentrator, sun shape, circumsolar ratio,
acceptance angle, efficiency, capacity factor, solar multiple, levelized cost
of energy.
2.1
Introduction
A concentrating solar power (CSP) system can be presented schematically
as shown in Fig. 2.1. All systems begin with a concentrator; the various
standard configurations of trough, linear Fresnel, dish and tower have been
introduced in Chapter 1, and are addressed in detail in later chapters. There
is a clear distinction between the line-focusing systems which concentrate
solar radiation by 50–100 times, and the point-focus systems that concentrate by factors of 500 to several thousand.
The concentrated radiation must be intercepted by a receiver which
converts it to another form, typically thermal energy. The currently dominant trough-based CSP systems use receivers that are single steel tubes
covered by a glass tube, with the annular space evacuated to reduce convection heat losses. Another commonly used option is to arrange multiple
tubes to form cavity shapes (either line- or point-focus). Alternatively,
‘volumetric’ or direct absorption receivers aim to have the radiation
absorbed by surfaces directly immersed in the working fluid. This can be
16
© Woodhead Publishing Limited, 2012
Fundamental principles of concentrating solar power (CSP) systems 17
Solar
radiation
in
Electricity
out
Direct connection
Concentrator
Receiver
Conversion
Energy
transport
Losses
Energy
storage
Heat to
environment
2.1 Schematic representation of the component parts of a solar
thermal power system.
done by having a window in front of a cavity containing an absorbing matrix
which the fluid passes over. Later chapters present details on possible
receiver types for the various concentrator technologies.
After the receiver, there are two options: either the energy is further
converted to the final form desired (such as electricity), or it is transported
to another location for final conversion. It is possible that the power cycle
is built integrally into the receiver unit (Stirling engines, for example). Solid
state (semiconductor material) conversion devices such as concentrating
photovoltaics and thermoelectric converters also lead to receivers built
from the devices themselves.
If power conversion is carried out remote from the receivers, the collected thermal energy is carried away in a heat transfer fluid (HTF). For
the trough plants built to date, this is predominantly a type of oil chosen
for its transport properties as well as thermal stability. Direct steam generation has been used with all concentrator types and has the advantage that
the HTF and power cycle working fluid are one and the same, eliminating
the need for a heat exchanger. Molten salt as HTF was pioneered in tower
systems and has also been introduced for troughs. It has the advantage that
the HTF is then also a favourable energy storage medium. Use of air as a
HTF has also been demonstrated, and chemical reaction systems are under
development as heat transfer mechanisms.
Choice of the transport path provides the option of thermal energy
storage (TES) in the intermediate thermal form before going to final conversion to electricity. The current commercially dominant approach is to
use molten salt in high temperature insulated tanks. Chapter 11 covers
thermal energy storage options in detail. There is also the option of designing an energy storage system after conversion to electricity; however,
© Woodhead Publishing Limited, 2012
18
Concentrating solar power technology
electricity storage approaches are not integral to the CSP system itself
but rather are independent systems that could be applied to any form of
electricity generation and are not addressed in this book.
The final stage in a CSP system is electric power generation. The dominant approach here is steam turbines, with Stirling engines, organic Rankine
cycles, Brayton cycles and photovoltaics also successfully proven. The efficiency of each subsystem can be defined as the ratio of energy out to energy
in. The overall solar-to-electric conversion efficiency for the CSP system
(ηsystem) is the product of the various subsystem efficiencies (concentrator/
optical, receiver, transport, storage and conversion):
ηsystem = ηoptical × ηreceiver × ηtransport × ηstorage × ηconversion
[2.1]
These can be considered at a particular instant or averaged over a timescale
such as a day or a year. Alternative naming of these efficiencies are frequently seen, and subsystems can be further grouped or subdivided according to what is being analysed.
The driving principles behind the development of CSP1 systems are
that:
•
•
•
final conversion of collected thermal energy to electricity is more efficient if the energy at the conversion subsystem is available at a higher
temperature;
countering this, energy losses from receivers increase with temperature,
but can be reduced by reducing the size of the receiver via concentration
of the radiation;
cost factors and material limits sometimes determine that the optimal
operating conditions must be lowered.
This chapter reviews the various fundamentals that contribute to these
principles and lead to the design of systems that seek to maximize overall
conversion efficiencies. The chapter ends with an introduction to the key
aspects of the economic analysis of CSP, since ultimately it is the cost of
production of energy that matters most. Cost of energy depends strongly
on the installed cost per unit of generating capacity plus also the level of
solar resource and financial parameters. It is thus affected by both the efficiency of systems and their cost of construction. Ultimately the design
process is one of ‘thermo-economic’ optimization (see, for example, Bejan
et al., 1996).
1
This discussion applies most directly to CSP systems with thermal energy conversion; concentrating photovoltaic (CPV) systems are discussed briefly in Section 2.7.5 and then further
in Chapter 10.
© Woodhead Publishing Limited, 2012
Fundamental principles of concentrating solar power (CSP) systems 19
2.2
Concentrating optics
2.2.1
Solar radiation
To a good approximation, the sun is a black body2 radiation source at its
surface temperature of 5,760 K (5487°C). By the time it reaches the earth’s
surface, the solar spectrum has been selectively absorbed at various wavelengths by the various constituents of the atmosphere, and 5,200 K black
body behaviour is a good approximation. Further details on the solar spectrum can be found in Chapter 3.
Concentrating systems only make use of the directly radiated component
of solar radiation; they do not collect diffuse or scattered radiation. Direct
normal irradiance (DNI) is the flux density of direct (un-scattered) light
from the sun measured on a flat plane perpendicular to the sun’s rays.
Insolation, radiant flux, flux density and irradiance3 are terms that are used
fairly interchangeably in solar technology discussions, for the rate of solar
radiation energy flow through a unit area of space, with SI units of W/m2
(symbol G). The solar constant, which is the intensity on a plane outside
the earth’s atmosphere, is approximately 1,367 W/m2 (Kopp and Lean,
2011). The maximum terrestrial DNI varies significantly with location and
weather conditions, but is often taken as 1,000 W/m2.
The heart of a CSP system is its mechanism for concentrating the solar
radiation to higher intensities. An important metric is the concentration
ratio. Concentration ratio can be defined in several ways, with two in
common use.
•
The optical concentration ratio, Co, is the ratio of irradiance at the
receiver surface Gr to the incident solar irradiance G:
Co =
Gr
G
[2.2]
Co is defined at any point of an output flux distribution, with special
reference being given to the point of maximum light intensity and concentration ratio which occurs at the peak of a flux distribution.
• The geometric concentration ratio Cg, is the ratio of collector aperture
area Ac to receiver area4 Ar:
2
A black body is an ideal surface that absorbs all radiation incident on it and radiates a defined
spectrum and intensity of radiation according to its own temperature. See, for example,
Bergman et al. (2011).
3
Intensity is also sometimes used in discussing solar irradiation; however, strictly speaking,
intensity is not a radiant flux but radiant power per unit solid angle from a source.
4
The area that is referred to here is the useful active absorbing area, often defined by an
aperture.
© Woodhead Publishing Limited, 2012
20
Concentrating solar power technology
100
Normalized solar intensity
10–1
10–2
10–3
10–4
CSR 40%
CSR 30%
CSR 20%
CSR 10%
CSR 5%
CSR 0%
10–5
10–6
100
101
Angular displacement (mrad)
2.2 Sun shape (intensity versus angle, averaged across a large
dataset) as a function of circumsolar ratio (Buie et al., 2003).
Cg =
Ac
Ar
[2.3]
Cg is subject to the receiver area chosen for analysis. Receiver apertures
can be chosen by design to capture as much of the focal region radiation
as possible, or limited to capture only the more intense portion.
Concentration ratio values are often referred to in terms of a number of
suns: a geometric concentration ratio of 1,200 would, for example, be said
to be ‘1,200 suns’. At an assumed solar flux of 1,000 W/m2 this would mean
an average 1.2 MW/m2 at the receiver surface.
Viewed from the earth, the sun subtends a half-angle of approximately
4.65 mrad (milliradians) (Fig. 2.3). However, the exact half-angle is complicated somewhat because the intensity distribution at the edges of the sun
is not a clear step-function. Instead it falls off over a narrow angular range
as shown in Fig. 2.2. This distribution of solar intensity with angular displacement is commonly referred to as the sun shape.
A key parameter is the circumsolar ratio (CSR), defined (Buie et al.,
2003) as:
CSR =
Gcs
,
Gcs + Gs
© Woodhead Publishing Limited, 2012
[2.4]
Fundamental principles of concentrating solar power (CSP) systems 21
where Gs is the solar intensity integrated from just the solar disc, out to its
limit at 4.65 mrad, while Gcs is the solar intensity integrated over the
annulus from 4.65 mrad to the outer extent of the solar aureole (surrounding glow), taken as 2.5° (43.6 mrad) for the sake of easy measurement using
laboratory equipment.
High circumsolar ratios can significantly reduce the capture efficiency of
high concentration collectors, due to a larger fraction of flux spillage. As
the circumsolar ratio changes, the sun-shape distribution (intensity versus
angle) also changes (Fig. 2.2). The actual sun shape is most strongly influenced by prevailing atmospheric conditions, particularly the level of particulate matter or moisture in the air.
A number of different sun-shape distributions have been proposed; a
commonly used one is that by Buie et al. (2003) which gives:
⎧ cos (0.326 ⋅ θ )
, 0 ≤ θ ≤ θs
⎪
I r (θ ) = ⎨ cos (0.308 ⋅ θ )
⎪⎩θ γ expκ ,
θ > θs
[2.5]
where
γ = 2.2 ⋅ CSR 0.43 ⋅ ln (0.52 ⋅ CSR ) − 0.1
κ = 0.9 ⋅ CSR −0.3 ⋅ ln (13.5 ⋅ CSR )
and Ir(θ) is the relative solar intensity, relative to the intensity measured at
θ = 0°. Other sun-shape distributions used for modelling purposes include
the Kuiper sun shape (Buie et al., 2003), and that of Rabl and Bendt (1982).
The exact sun-shape distribution used for modelling purposes makes
little difference except when modelling concentrators have very high optical
concentration ratios (>10,000 suns), as in dish concentrators or solar furnaces, provided the optical surface is also very accurate. When dealing with
lower-concentration collectors, a flat-topped intensity distribution, often
called a pill-box sun shape is sufficient.
2.2.2
Calculation of sun position
A solar concentrator, whether line-focusing or point-focusing, needs to be
aligned to the direction of the incident solar rays. Sun position can be very
precisely predicted with a range of well-established equations as discussed
in Chapter 3.
2.3
Limits on concentration
Higher levels of concentration offer the benefit of reduced thermal
losses from smaller receiver apertures. However, there is a fundamental
© Woodhead Publishing Limited, 2012
22
Concentrating solar power technology
Solar
disc
le
ng
lf a
Ha
Reflected
cone
e
gl
f
al
an
H
Reflector surface
element
2.3 Points on a reflector surface reflect direct solar irradiation in a
cone of rays.
thermodynamic limit to achievable concentration. There are further limits
that result from particular concentrator geometries and then a range of
practical factors that limit it further again. An authoritative presentation of
the limits to concentration is given by Winston et al. (2005). In this section
an overview of the principles and results is given.
2.3.1
A limit from the second law of thermodynamics
As a result of the finite angular width of the sun, any element of a mirror
surface in a concentrator system will effectively reflect a cone of radiation
with the same angular spread, as illustrated in Fig. 2.3. Energy transfer
between the sun and the receiver of a solar concentrator is subject to the
second law of thermodynamics. This means that the solar receiver cannot
attain a higher temperature than that of the sun. Using this principle, limits
on the geometric concentration ratio can be established.
Consider the sun as a black body sphere of radius r a distance R from an
observer as shown in Fig. 2.4. At a distance R from the sun, all the radiation
leaving the surface will be uniformly distributed across a sphere of area
4πR2. Thus, the irradiance will fall off with distance according to:
⎛ ( R sin θ s )2 ⎞
⎛ r2 ⎞
2
G = E0 ⎜ 2 ⎟ = E0 ⎜
⎟⎠ = E0 sin θ s ,
⎝R ⎠
R2
⎝
[2.6]
E0 = σTs4
[2.7]
where
is the black body emissive power from the sun’s surface, and σ = 5.670 ×
10−8 W/m2K4 is the Stefan-Boltzmann constant.
© Woodhead Publishing Limited, 2012
Fundamental principles of concentrating solar power (CSP) systems 23
r
qs
R
Sun
Earth
2.4 Radiation flux from a spherically symmetric black body falls off as
1/R2.
q
Focused radiation
of angular spread
90°
Incoming radiation
of angular spread q
2.5 A concentrator that takes radiation with angular spread half-angle
θ and concentrates it to a receiver with a final angular spread of halfangle 90°.
Now, imagine some ideal solar concentrator that takes solar radiation
with angular spread5 θ and accepts it from throughout a certain collector
aperture area Ac, concentrating it onto a black-body receiver of some area
AR in a manner such that at the point of incidence, the angular spread has
a half-angle of 90° (Fig. 2.5). The black-body receiver will heat up and all
of its diffusely emitted radiation will follow a reverse path out of the concentrator and back to the sun. In the absence of any other heat losses, the
black-body absorber will heat up until it reaches the same temperature as
the source, and it will then be in equilibrium. This implies that
σTs4 Ac sin 2 θ = σTR4 AR,
[2.8]
where TR is the absolute temperature of the receiver and Ts is the temperature of the sun’s surface. Since TR = TS at equilibrium,6 this gives the result
The angle θ is now generalized to be the acceptance angle which could be more or less
than θs.
6
Having established this limit under the condition of thermal equilibrium, we can then argue
that it also applies away from equilibrium, since the reflections and refractions that lead to
the optical concentration at the receiver are not a function of the receiver temperature. Hence
this limit on concentration ratio is general.
5
© Woodhead Publishing Limited, 2012
24
Concentrating solar power technology
f
q
Ain
Incoming radiation
of angular spread q
Aout
Outgoing radiation
of angular spread f
2.6 An arbitrary concentrator accepting radiation with a half-angle θ
over area Ain and sending it out over area Aout with half-angle φ.
(the sine law of concentration) that any point-focus solar concentrator must
have a concentration ratio of no more than
Cg =
Ac
1
.
=
AR sin 2 θ
[2.9]
Going beyond this, imagine that the planar black-body receiver is
replaced by an aperture, leading to a further optical system that transforms
the radiation to a new angular spread and a third aperture area (Fig. 2.6).
For the second optical subsystem, the aperture appears as a black body also
and the same arguments apply leading to
σTR4 AR = σTs4 Aout sin 2 φ.
[2.10]
The whole device can now be viewed as an arbitrary ideal optical concentrator with
σTs4 Ain sin 2 θ = σTs4 Aout sin 2 φ
[2.11]
and hence,
C g 3D =
Ain
sin 2 φ
=
.
Aout sin 2 θ
[2.12]
For a solar concentrator, θ is interpreted as the acceptance angle of the
concentrator and φ as the angular spread of the concentrated radiation
incident on the receiver.
Thus, the second law of thermodynamics leads to the conclusion that the
concentration of radiation can only be achieved by increasing its angular
spread, and this inherently limits the extent to which non-parallel radiation
can in fact be concentrated.7
7
This result is known in non-imaging optics as the principle of conservation of étendue. A
rigorous proof is given by Chaves (2008).
© Woodhead Publishing Limited, 2012
Fundamental principles of concentrating solar power (CSP) systems 25
A more rigorous analysis considers the possibility of materials with
different refractive indices nin and nout at the entrance and exit respectively,
and gives the result
2
⎛ n sin φ ⎞
C g 3 D = ⎜ out
.
⎝ nin sin θ ⎟⎠
[2.13]
however, for most CSP purposes, concentration takes place in air, so
nin = nout.
Note that the concentration depends on the acceptance angle of the
concentrator, rather than the actual angular spread of the light source. A
collector could be made with a very high concentration ratio by choosing
to have a very narrow acceptance angle, even narrower than the angular
spread of sunlight, but it would not help greatly, since it would then only
collect a small fraction of the solar radiation. Conversely a wide acceptance
angle could be chosen for a non-tracking concentrator at the expense of
low achievable concentration ratio. Typically, the acceptance angle of a solar
concentrator will be fairly close to the angular size of the sun.8
If the analysis is repeated for a line-focus concentrator, then the
geometric limitations on acceptance apply only in one direction, and the
flux density falls off as 1/R rather that 1/R2, giving the result that
Cg 2D =
nout sin φ
.
nin sin θ
[2.14]
Using a refractive index of 1 for air, and an exit half-angle φ = 90° for
maximum concentration, brings the result for a line-focus concentrator to
C g 2 D max =
1
.
sin θ
[2.15]
For linear concentrators, once again it is the acceptance angle that matters
and this is effectively 90° in the plane of the sun and linear receiver, so
concentrating light from the sun is still limited by Eq. [2.15] even though
it is actually a spherical source. These final results (Eqs [2.9] and [2.15])
limit the maximum geometric concentration ratio of any solar
concentrator.
Evaluating these limit equations for the actual angular size of the
sun (half-angle 4.65 mrad) shows that the thermodynamic limit concentration for point-focus concentrators is 46,250 and for linear concentrators
215.
8
An exception is the compound parabolic concentrator (CPC) used in certain non-tracking
concentrators, such as for backing reflectors in solar hot water systems. CPCs also play a role
in the secondary optics in certain types of systems (Meinel and Meinel, 1976).
© Woodhead Publishing Limited, 2012
26
Concentrating solar power technology
Focal plane
y
Focus
Axis
fR
Rim angle
fR
f
Vertex
ZR
x
W
2.7 The parabola has the property that, as a reflector, all incident rays
parallel to the axis will be reflected to pass through a single point at
the focus.
2.3.2
Parabolas and paraboloids
The central role of the parabola in solar concentrators stems from its ability
to focus parallel radiation to a point a distance f from its vertex. The
parabola is a two-dimensional curve that provides the cross-sectional shape
of a trough-shaped linear concentrator. Similarly, the paraboloid is a
parabolic surface of revolution and is the surface shape given to dish
concentrators.
The functional relationship defining a parabola with its axis aligned to
the y-axis (Fig. 2.7) is
y=
x2
,
4f
[2.16]
while, for a paraboloid with its axis coinciding with the z-axis, it is
z=
r2
x 2 + y2
.
=
4f
4f
[2.17]
The rim angle φR is the angle between the axis and a line from the focus
to the physical edge of the concentrator. Together, the focal length and rim
angle of a parabolic concentrator completely define its cross-sectional
geometry. The rim angle of a parabola or paraboloid is given by
tan φR =
W/ 2
4 fW/ 2
=
f − zR 4 f 2 − (W/ 2)2
where W is the width and zR is the depth of the parabola at the rim.
© Woodhead Publishing Limited, 2012
[2.18]
Fundamental principles of concentrating solar power (CSP) systems 27
qs
qs
y
d/2
r sin qs
fR
qs
r
f
x
W
2.8 Concentrating solar radiation with a perfect parabolic mirror to a
flat target.
The parabolic effect of focusing to a single point only occurs with perfectly parallel incoming radiation. As noted above, sunlight has a range of
incident angle due to the physical extent of the solar source. This reduces
the optical concentration achievable at the focus of parabolas and paraboloids, as discussed in the following sections.
Limits with flat receivers
Each point on a parabolic mirror will reflect a cone of rays that matches
the angular distribution of the solar source (half-angle size θs). Consider
the size of the spot formed by the cone of rays reflected from the points
on the mirror, when incident on a flat target placed in the focal plane
(Fig. 2.8). The rays from the rim will form the widest such spot.
The distance x of the reflection point from the axis is
x = 2r sin φR,
[2.19]
9
and the width of the focal spot on the focal plane due to reflection from
this point will be
d=
2r sin θ s
.
cos φR
[2.20]
9
Consideration of a precise width in this way is based on an assumed pill-box sun shape, an
approximation.
© Woodhead Publishing Limited, 2012
28
Concentrating solar power technology
The reflected cone from a single spot on the mirror will actually form an
elliptical spot on the target with a major axis length of d. This applies
equally for dishes and troughs.
Clearly the size of the focal-plane spot depends on the received incidence
angle φ, so the accumulated effect of all the spots across the mirror will not
be a true image of the sun formed at the focal plane. This is the reason that
solar concentrator optics are referred to as non-imaging optics. If the rim
angle φR is small, then r is approximately equal to f over the whole mirror,
and then, to a good approximation, it will form a true image of the sun in
the focal plane,10 with the image diameter being
d=
2 f sin θ s
= 2 f sin θ s .
cos 0
[2.21]
On the other hand, for a dish or trough with a flat receiver and a rim angle
of 90°, the mirror elements at the very edge will make an infinitely long
spot, spreading their radiation over the entire focal plane.
If the receiver is large enough to accept reflected spots from the entire
mirror surface, then the diameter of the receiver will be defined by the
reflected spot size from points at the edge of the mirror with x = W and
φ = φR. The geometric concentration ratio for a parabolic trough with
flat receiver will then be
Cg =
AC LW W
=
=
.
AR
Ld
d
[2.22]
After substituting Eqs [2.19] and [2.20], this gives
Cg =
sin 2φR
.
2 sin θ s
[2.23]
To find the optimal rim angle for a parabolic trough, take the derivative
with respect to φR
dC g
d
=
dφ R dφ R
⎛ sin 2φR ⎞ 2 cos 2φR
= 0 at φR = 0, 45°, 90°, . . .
⎜⎝
⎟=
2 sin θ s ⎠
2 sin θ s
[2.24]
The maximum concentration ratio corresponds to φR = 45°, and gives a
maximum concentration ratio11 for a trough, with flat receiver and solar
acceptance angle θs = 4.65 mrad, of
C g ,trough, flat,max =
1
≈ 108.
2 sin θ s
10
[2.25]
This is the case with astronomical telescopes; they cannot have a large rim angle or else their
imaging quality will be lost.
11
This analysis ignores the effect of mirror shading by the receiver, and also considers only a
single-sided receiver. For a fuller treatment, see Rabl (1976).
© Woodhead Publishing Limited, 2012
Fundamental principles of concentrating solar power (CSP) systems 29
Repeating this analysis for a paraboloidal dish
π
2
2
W2
W
Ac
⎛ sin 2φ ⎞
,
Cg =
= 4π 2 = ⎛⎜ ⎞⎟ = ⎜
⎟
⎝ d⎠
⎝ 2 sin θ s ⎠
AR
d
4
[2.26]
from which
dC g
d ⎛ sin 2 2φ ⎞ sin 2φ cos 2φ
=
=
dφ
dφ ⎜⎝ 4 sin 2 θ s ⎟⎠
sin 2 θ s
= 0 again gives maximum at φ = 45°.
[2.27]
Hence
C g ,dish , flat ,max =
1
≈ 11, 600.
4 sin 2 θ s
[2.28]
These values are a half (trough) and a quarter (dish) of the thermodynamic
limits of Section 2.3.1. Most dish and trough concentrators with cavity
receivers (which have a flat opening) employ a rim angle close to 45°.
Limits for cylindrical and spherical receivers
Another possibility to consider is using a receiver with a circular crosssection as shown in Fig. 2.9. In this case, the diameter of the target needs
to be
d = 2r sin θ s .
[2.29]
qs
qs
y
r sin qs
fR
qs
r
f
x
W
2.9 Concentrating solar radiation with a perfect parabolic mirror to a
circular cross-section target.
© Woodhead Publishing Limited, 2012
30
Concentrating solar power technology
For a trough with cylindrical receiver, then
Cg =
sin φ
Ac
LW
.
=
=
AR L2π r sin θ s π sin θ s
[2.30]
Solving for maximum geometric concentration ratio as before, the optimal
trough rim angle is φ = 90°, and that at this angle, gives
C g ,trough,cyl ,max =
1
≈ 68.5.
π sin θ s
[2.31]
Trough concentrators with cylindrical evacuated-tube receivers consequently employ rim angles approaching12 φ = 90°.
For a dish with a spherical receiver, likewise,
Cg =
π
π
4r 2 sin 2 φ
W2
sin 2 φ
Ac
4
4
,
=
=
=
AR 4π r 2 sin 2 θ s 4π r 2 sin 2 θ s 4 sin 2 θ s
[2.32]
and again the maximum geometric concentration ratio occurs at φ = 90°,
with
Cg =
1
≈ 11, 600.
4 sin 2 θ s
[2.33]
For troughs then, the optical analysis gives a limit that is equal to the thermodynamic limit divided by π; for dishes, the result is equal to one quarter
of the thermodynamic limit.
Note that this analysis is of geometric concentration ratio, and the above
derivation indicates that the local contributions to focal spot size vary as a
function of reflection radius x, so the geometric concentration ratio limits
are less than the optical concentration ratio limits at the centre of the focal
spot.
Rabl (1976) has further results giving the mirror area per aperture area
for these four different collector configurations. Optimally sized troughs
and dishes with flat receivers require less mirror for a given aperture than
those with cylindrical/spherical receivers, since on average the glass is
reflecting at closer to a normal angle.
2.3.3
Secondary optics
The addition of a second stage of optical concentration, or secondary optics,
can lead to higher concentration ratios with real results up to 80–90% of
the thermodynamic limit achieved (Gordon, 2001). The single-stage concen12
In practice, lower angles of the order of 80° are commonly used, because of increasing cosine
losses per glass area and increasing amplification of surface errors due to the great distance
from mirror to receiver.
© Woodhead Publishing Limited, 2012
Fundamental principles of concentrating solar power (CSP) systems 31
2.10 A secondary Trombe-Meinel cusp concentrator, shown here in
the Novatec linear Fresnel system, can allow concentration ratios on
the absorber surface to further approach the thermodynamic limit.
This receiver is also discussed in Chapter 6.
trators described above often have incident radiation at their receiver from
only a relatively limited angular range (for flat absorber, a half-angle of
45°). Conservation of étendue suggests, then, that further concentration is
possible; this is achieved in practice using various funnel-like mirror systems.
The compound parabolic concentrator (CPC, also Winston collector), the
Trombe-Meinel cusp (Fig. 2.10) and the Mouchot conical mirror are geometric configurations that can be used to gain this further degree of
concentration.
2.3.4
Practical factors reducing concentration
A range of sources of errors in solar concentrators act to reduce the concentration that can be achieved. They can be categorized according to the
scale of the imperfection causing the error, starting at the scale of microns,
moving up to the scale of centimeters.
Specularity error
Specularity refers to the mirror-like quality of a reflector, or specifically, the
degree to which reflected rays obey the law of reflection, where the reflected
angle equals the incident angle. The opposite is a diffuse reflector, which
scatters reflected light in a wide range of directions.13 The specularity error
13
A special case is the Lambertian surface, for which the apparent brightness of reflected
radiation is equal in all directions. Such surfaces are useful in methods for characterizing
concentrator performance, since they allow a photograph to be taken to record the irradiance
in the focal plane of a dish, heliostat, etc.
© Woodhead Publishing Limited, 2012
32
Concentrating solar power technology
σsp is defined locally on a surface as the standard deviation of the distribution of reflected ray angles at a specified incident angle; all real surfaces
have some degree of specularity error, arising from the microscopic properties of surfaces.
Surface slope error
At a slightly larger scale, concentrator mirrors will have local ripples and
distortions in their surfaces, and the degree of aberration is often called the
surface slope error σsurf, defined as the standard deviation of the angular
deviations of the surface normal vectors from their ideal directions, sampled
across the surface of the mirror.
The larger the surface slope error, the poorer the optical performance of
the concentrator: the focal spot spread increases, and the maximum concentration ratio decreases. Values of around 0.4 mrad are found in very
accurate systems, although values up to around 5 mrad can be adequate,
depending on the type of concentrator and application.
Shape error
Looking at a larger scale, a concentrator is commonly constructed of facets;
these facets may be oriented incorrectly, or there may be distortion of the
overall structure due to thermal expansion, wind loads or the release of
residual stresses. Using accurate measurement techniques such as photogrammetry, these errors can be measured and converted into an effective
shape error σsh, which is the standard deviation in the surface normal angles
arising from these various forms of misalignment.
Tracking error
A tracking system ideally should make the concentrator point at the sun
without error. In reality, tracking systems are not perfect, and will not
always point the collector exactly at the sun. This angular offset often varies
with time, particularly with an ‘on-off’ type tracking control system. When
considered over a period of time greater than the source of the error, the
angular error can be characterized by the tracking error σtr, defined as the
standard deviation of angular error distribution.
Combinations of errors
In a typical system, the specularity, slope, shape, and tracking errors can be
combined into an overall optical error, σtot, of a concentrator. If all the
sources can be modelled to reasonable accuracy as a normal distribution of
© Woodhead Publishing Limited, 2012
Fundamental principles of concentrating solar power (CSP) systems 33
randomly distributed errors, then the overall optical error is found as a root
sum of squares of the component errors. That is,
2
2
2
σ tot = σ sp
+ σ surf
+ σ sh
+ σ tr 2 .
[2.34]
Certain types of surface and shape error may not be well characterized by
a normal distribution, so this equation should be used with care. Also,
if modelling the sun shape as a simplified normal distribution, then an
additional term σ2sun can be added inside the square root of the above
equation.
2.3.5
Cosine losses and end losses
An important source of ‘loss’ in solar concentrators arises from the fact that
mirrors cannot always be aligned normal to the incident solar rays. When
a mirror is reflecting off-axis, the apparent area of the mirror, as seen from
the sun, is reduced according to the cosine of the incidence angle. Assuming
that the aperture area of the concentrator to be equal to the mirror area,
this reduction in apparent area then directly reduces the concentration ratio
of the concentrator, hence it is referred to as a cosine loss, although strictly
speaking the energy was never collected in the first place.
This cosine loss effect occurs in all forms of solar concentrators. For all
concentrators except for dishes or lenses, cosine losses vary as a function
of sun position. Dishes are always pointed directly at the sun, so the only
cosine losses arise from the curvature of the mirrors. An attempt to minimize average cosine losses is responsible for the off-axis arrangement of
most central receiver systems.
End losses are particular to trough and linear Fresnel concentrators.
These refer to the radiation that is reflected from the mirrors but which,
due to the sun not being directly overhead of the collector, misses the
receiver, and instead is concentrated beyond the end of the receiver.
Depending on the orientation (east–west or north–south), these losses
might be present all year round, or else only in the mornings and
evenings.
Both end losses and cosine losses are commonly incorporated as part of
the calculation of ηoptical in Eq. [2.1].
2.4
Focal region flux distributions
Once light has been reflected or refracted on its path through a concentrator, there will be a distribution of irradiance in the focal region. Predicting
and measuring the focal region flux distribution is essential for design and
performance analysis of receiver systems.
© Woodhead Publishing Limited, 2012
34
Concentrating solar power technology
2.4.1
Prediction of focal region distributions
For the prediction of the focal region distribution, a common technique is
ray tracing. This computational technique involves gathering information
about the collector geometry, together with the sun shape, specularity,
slope, shape and tracking errors, and then creating a virtual ‘scene’ containing all of those elements and their associated parameters. Large numbers
of simulated rays are then ‘fired’ into the virtual ‘scene’, with an angular
distribution that matches the specified sun shape. Each ray is traced
through the scene. At an obstacle, some of the ray’s strength will be attenuated according to the surface reflectivity and the direction of the reflected
ray will be assigned based on a random sample from the slope error distribution and the specularity distribution. The process of ‘tracing’ is
repeated until the ray leaves the scene or is completely absorbed. Refractions through glass, etc., can also generally be accommodated. A grid of
cells is defined on various ‘virtual sensor surfaces’ in the scene (usually
corresponding to a receiver surface or the focal plane). Then radiation
distributions can be calculated, by summing the strength of all the rays
that are incident on each grid cell of the sensor surface. Accuracy is
increased by calculating with larger numbers of rays and using smaller grid
cells, but at the expense of increased calculation time. Several million rays
are frequently required to achieve sufficient accuracy in the analysis of a
solar concentrator.14
Leading general-purpose ray-tracing software used in the CSP field
includes ASAP, OptiCAD, OSLO, and ZEMAX. A recent entrant is Tonatiuh, a free open-source ray-tracer specifically for solar concentrators, under
active development at CENER in Spain (Blanco et al., 2011). Some optical
codes specifically developed for solar applications include DELSOL,
MIRVAL, UHC, HFLCAL, FIAT LUX and SolTrace (Garcia
et al., 2008; Ho, 2008). Some of these codes make use of a convolution
approach, which treats the distribution of rays as a whole, rather than calculating the path of each individual ray through the scene; the approach is
reported to be within 1–2% of detailed ray-tracing, but significantly faster
to calculate (Garcia et al., 2008).
2.4.2
Measurement of focal region distributions
Focal region distributions can be measured directly using the various
methods discussed in detail in Chapter 18. The most common of these is
14
Note that CSP ray-tracing is quite distinct from the ‘backward’ ray-tracing used in CGI,
animated movies, etc., which considers possible origins/luminous intensity of rays emanating
from the eye of the observer in different directions.
© Woodhead Publishing Limited, 2012
Fundamental principles of concentrating solar power (CSP) systems 35
250
200
150
100
50
0
0
120 00
10 00
8
600
400
200 0
0
600
400
200
800
1000
1200
2.11 Experimentally determined irradiance distribution of the ANU
500 m2 dish at the focal plane. Spatial units are millimetres, the
vertical scale uncalibrated relative units (Lovegrove et al., 2011).
the camera target method. A suitably cooled Lambertian target is placed
at the focal plane of the concentrator. Then, using a camera, images of the
reflected radiation from the target are captured. If the reflectivity of the
surface is known and the camera is well calibrated, then the irradiance at
the target can be determined.
An example of an experimentally determined distribution from the ANU
500 m2 dish system (Lovegrove et al., 2011), showing a typical Gaussian-like
distribution, is shown in Fig. 2.11.
If a real receiver geometry is superimposed on a known focal region
distribution, the fraction of the solar radiation initially intercepted by the
concentrator aperture that is in turn intercepted by the receiver can be
determined. This capture fraction or intercept factor is a major determinant
of the optical efficiency of the system. Figure 2.12 illustrates the intercept
fraction of a circular aperture imposed on the flux distribution of Fig. 2.11
as a function of diameter.
Optical errors lead to a spreading of the focal region distribution, and a
reduction in the achievable concentration ratio. This leads to a need for
larger receivers, so as to avoid flux spillage, or in other words rays that miss
the receiver and are lost. On the other hand, an analysis of losses reveals
that having an excessively large receiver will result in large radiative and
convective losses, so, in general, some level of flux spillage will be tolerated
for an optimized collector.
Knowledge of the flux density distribution allows a collector efficiency to
be determined based on the actual amount of flux intercepted by the
receiver aperture compared to that intercepted by the collector. This could
© Woodhead Publishing Limited, 2012
36
Concentrating solar power technology
Percent capture
SG4 Full moon 4 Sep 2009 - image 35
100
90
80
70
60
50
40
30
20
10
0
0
0.1
0.2
0.3
0.4
Diameter (m)
0.5
0.6
0.7
2.12 Empirical relative intensity distribution of the ANU 500 m2 dish at
the focal plane (Lovegrove et al., 2011).
be evaluated at a particular instant or averaged over a time interval to
capture changes in tracking or structure; for example
ηoptical ≡
∫ ∫
Gincident (t ) dA dt
time Aperture
∫
Acollector Gsol (t ) dt
[2.35]
time
where Gsol(t) is the time varying DNI and Gincident(t) is the time varying
concentrated irradiation at the receiver and Acollector is the aperture area of
the collector.15 If irradiation is constant over the time interval in question,
then the time integrals are no longer needed.
This definition captures several effects:
•
•
•
•
the actual capture fraction for a perfectly aligned concentrator for a
given receiver
the loss due to non-unity mirror reflectivity
the loss due to tracking error
cosine losses.
2.5
Losses from receivers
At steady state, the net energy flow into a receiver from concentrated solar
radiation will be balanced by energy outflows from the flow of heat transfer
15
Note that evaluations can sometimes be presented based on aperture with or without the
effect of receiver shading or total mirror area rather than aperture area, so results must be
interpreted carefully.
© Woodhead Publishing Limited, 2012
Fundamental principles of concentrating solar power (CSP) systems 37
fluid or other energy conversion process, plus a range of energy losses, due
to unwanted reflection, radiative emission, convective or conductive processes, detailed in the following sections.
The total loss will be the sum of these four contributions
Q loss = Q ref + Q rad + Q conv + Q cond
[2.36]
The energy efficiency of a receiver is the ratio of the energy that is usefully
converted to the input energy from the concentrated that is intercepted
ηrec =
Q converted Q input − Q loss
=
Q input
Q input
[2.37]
where for the receiver
Q input =
∫ ∫
Gincident (t ) dA dt
[2.38]
time Aperture
If the useful energy conversion process is ‘turned off’, the receiver efficiency
will be zero and the receiver will heat up until the combined losses exactly
balance the incident radiation. This temperature is referred to as the stagnation temperature.16
2.5.1
Radiative losses
Radiative loss processes include both the net emitted radiation from receivers as a consequence of their temperature and the reflection of some of the
incident concentrated solar radiation. Surfaces emit and absorb radiation
as essentially independent processes, with the net energy transfer taking
place being the combination of the two. Each surface in a receiver will emit
radiation in proportion to the fourth power of temperature, at a rate given
by the black body emissive power multiplied by its emissivity.
Some fraction of the incident radiation will be reflected from any surfaces
on which it is incident. The fraction of this radiation that is lost from the
receiver will depend on the geometry. If glass covers are involved in the
receiver construction, they will also introduce reflection losses that may be
mitigated by anti-reflective coatings. For cavity receivers in particular, radiation reflected or emitted from one part of a receiver surface is quite likely
to be incident on other parts, so calculating the net absorbed radiation
requires a simultaneous solution of the whole process. Ray-tracing software
typically does this for reflected incident radiation (but not emitted
radiation).
16
For low concentration systems, this is an experiment that can be performed, but for high
concentration point focus systems, destruction of the receiver is likely before an empirical
stagnation temperature could be established.
© Woodhead Publishing Limited, 2012
38
Concentrating solar power technology
Many real surfaces can be modelled as grey bodies, meaning that they
have a constant emissivity across all wavelengths of interest. An important
exception is the selectively absorbing surface which, in solar thermal applications, is designed specifically to have a high absorptivity in the wavelength
range of solar radiation and a low emissivity in the wavelength range associated with the (infrared) radiation emitted from hot receiver surfaces.
Chapter 15 covers such surfaces in detail.
For diffusely reflecting surfaces, the methods for calculating the final
distribution of incident solar radiation are the same as those required to
determine how the diffuse emitted radiation from hot receiver surfaces is
ultimately distributed.
The radiation leaving a surface will be partly intercepted by all the other
surfaces in the field of view, in proportion to the view factor (also called
radiation shape factors),
Fij ≡ the fraction of radiation leaving surface i and
reaching surface j.
[2.39]
If these other surfaces reflect and absorb various fractions, working out the
final distribution of absorbed energy becomes a complex problem. General
presentations of radiation heat transfer can be found in Bergman et al.
(2011).
The radiative energy balance of a particular (diffusely reflecting) surface
element is illustrated in Fig. 2.13. In this diagram, Gi is the total irradiance
Emitted
radiation
Diffusely
emitting/
reflecting
surface
Reflected
radiation
Absorbed
radiation
Total outgoing
radiation (Ji)
Divided
according to
Fij
Total incoming
radiation (Gi)
Composed
of
(1/Ai)SJjAjFji
2.13 Radiation energy balance on a diffusely emitting and reflecting
surface.
© Woodhead Publishing Limited, 2012
Fundamental principles of concentrating solar power (CSP) systems 39
incident on surface i, Ji is the radiosity of surface i, defined as the total
radiative flux (reflected plus emitted) leaving that surface, Fij is the view
factor from surface i to surface j, and Ai is the area of surface i. Each wavelength range (e.g. solar vs. thermal wavelengths) can be considered independently in this way.
For each surface in a radiation exchange network, an equation for the
energy balance is written. Together, these equations form a linear system in
terms of radiosity, and so they can be readily solved to evaluate the net heat
transfer at each surface. Boundary conditions are required at each surface;
the surface temperature can be fixed and the equation solved for net heat
transfer, or else vice versa.
For a solar concentrator receiver, the radiation exchange between surfaces within it can be solved in this manner. The starting point is that the
amount of concentrated solar radiation coming in through the aperture and
striking each surface needs to be known from the optical properties of the
concentrator. The aperture itself can be regarded as a black body surface
to all radiation incident on it from other internal surfaces. Boundary conditions for external convective losses can also be introduced, relating surface
temperature and the externally convected flux.
Analysis of simple grey-surface radiation exchange can be achieved with
the free open-source software View3D (Walton, 2002). More sophisticated
models incorporating coupled radiation and convection heat transfer are
described in Section 2.5.2.
In a simplified model, if the aperture is treated as a single surface at the
average receiver temperature, interacting only with the environment, then
the emitted radiation loss will be given by
4
4
− Tenv
Q rad = σ Aε FRS (Trec
),
[2.40]
where FRS is a simplified shape factor between receiver and surroundings.
In a simplified model, reflection losses may be approximated using a single
effective net absorptivity for the receiver aperture area. Thus
Q ref = (1 − α ) AQ sol .
2.5.2
[2.41]
Convection losses
Convection losses arise in solar concentrators from the movement of air
over hot receiver surfaces. Efforts can be made to minimize such losses,
such as by placing a glass cover over the receiver surface or using an evacuated glass tube (as with parabolic troughs). Some dish and tower systems
are fitted with quartz windows able to withstand very high solar flux. In
other cases, the receiver is an open cavity, and only the buoyancy of hot air
© Woodhead Publishing Limited, 2012
40
Concentrating solar power technology
trapped inside the downward-facing receiver acts to suppress the convective loss.
Convection losses, both forced convection due to wind and natural convection due to the buoyancy effects of heated air, are difficult to measure
and difficult to model. One complication is that the process of convection
is coupled with the process of radiative loss, in the sense that the magnitudes
of both types of loss are dependent on the surface temperature, which is in
turn dependent on the external (radiation and convection) heat loss as well
as the internal heat transfer (to the fluid passing through the inside of the
receiver). Efforts are often made to decouple the problem, such as by
assuming a uniform receiver wall temperature, or assuming uniform heat
flux into the heat transfer fluid.
Convection losses are very difficult to measure directly, and usually
empirical results are obtained by subtracting other known losses from the
overall energy balance. Detailed convection heat transfer simulations are
possible with commercial computational fluid dynamics (CFD) tools such
as FLUENT and Star-CCM+. A free open-source alternative is OpenFOAM
(OpenFOAM Foundation, 2011). For simplified analysis semi-empirical
correlations for natural and forced convection heat transfer coefficients17
can be used, but the accuracy will usually be no better than ±20%.
Once correlations have derived a value of the average convection heat
transfer coefficient (h), the receiver convective loss can be calculated
Q conv = hA (Trec − Tamb ) .
2.5.3
[2.42]
Conduction losses
There will be thermal losses through insulating covers over the back of
receivers and any thermal path between hot receiver surfaces and the surrounding environment. Such losses are also linearly proportional to the
temperature difference and inversely proportional to an overall ‘thermal
resistance’ that depends on material conductivities and geometry
(T − Tenv )
.
Q cond = rec
Rth
[2.43]
For a one-dimensional geometry of heat loss through a single homogeneous (insulating) layer of thickness L and thermal conductivity k, the
equation is
(T − Tenv )
.
Q cond = kA rec
L
17
[2.44]
Usually a dimensionless heat transfer coefficient (Nusselt number) is correlated with
Grashof and Prandtl number for natural convection, or Reynolds number and Prandtl number
for forced convection.
© Woodhead Publishing Limited, 2012
Fundamental principles of concentrating solar power (CSP) systems 41
2.6
Energy transport and storage
Energy transport is essentially about moving high temperature HTFs
through piping networks. Chemical reaction-based energy transport is also
under development. HTF networks can be extensive for distributed collector fields such as trough systems, whilst they are considerably smaller for
tower systems. The basic approaches to calculating pressure drop and
pumping power, plus heat loss through insulated pipes are well established
in engineering practice and are not discussed further here. It is worth noting,
however, that there are standard approaches for thermo-economic optimization of pipe diameter and insulation thickness that should be applied, but
noting that what is cost effective in standard practice may not transpire to
be so in a CSP system.
Energy storage is discussed in detail in Chapter 11. Key categories of
energy storage for CSP systems include sensible storage (heating and
cooling a material without change of phase), latent heat storage (melting
and freezing of suitable high-temperature phase-change materials) and
thermochemical storage (with reversible chemical reactions used to store
and discharge energy).
Two-tank systems, molten-salt systems, incorporating sensible heat transfer with hot and ‘cold’ tanks of molten NaNO3-KNO3, are currently the
commercially dominant approach. Most other approaches under development also involve storage of high-temperature material. Reducing direct
thermal conduction losses and parasitic energy consumption are obviously
important. A less obvious, but potentially more important, issue is that of
the temperature decrease that is experienced in directing thermal energy
to storage and then later extracting it. This temperature decrease results in
a direct loss of exergy, as discussed in Section 2.8.1.
2.7
Power cycles for concentrating solar power
(CSP) systems
A range of different solar to electric energy conversion systems can be
applied to the various concentrator types.
2.7.1
Steam turbines
The bulk of the world’s electricity is generated with steam turbines, mostly
with steam produced from fossil or nuclear heat sources. One of the advantages of CSP is the ease with which a new source of heat can be applied to
the dominant power generating technology. Consequently the vast majority
of the CSP systems presently in operation use steam turbines. All the
concentrator types have been applied to steam production for use in steam
turbine energy conversion.
© Woodhead Publishing Limited, 2012
42
Concentrating solar power technology
A plant with a Rankine cycle using a steam turbine works by:
• compressing pure feedwater to high pressure (over 10 MPa, for example);
• boiling and superheating steam in a boiler which may be in the focal
point, or may be heated using a heat exchanger with another heat transfer fluid;
• expanding the steam to low pressure via a series of turbines that drive
a generator; and
• at the end of the expansion process, condensing the low pressure steam
with the aid of a cooling tower and then re-using it in the cycle.
The Rankine cycle has higher conversion efficiency for higher steam temperature and pressure at turbine entry (in common with all heat engine
cycles). A key feature that improves efficiency is including various stages
of steam bleed from the turbines that can then be used to progressively
heat feedwater prior to use in boilers.
The fraction of liquid condensing within the turbine must be kept to a
low value to avoid blade erosion. This can be achieved by ensuring that the
vapour is sufficiently superheated prior to expansion. Increasing boiler
pressure to increase efficiency can mean that the materials will not allow
vapour to be superheated far enough to avoid condensation in the turbine.
This problem is addressed by re-heating the vapour after partial expansion.
All of these features are typically combined in a large-scale steam turbinebased power plant and the overall configuration is typically some variation
of that shown in Fig. 2.14.
Re-heat
Multi-stage turbine
Shaft power
to generator
High
pressure
Low
pressure
Heat
in
Boiler
Bled steam
Condenser
Closed fw
heater
Closed fw
heater
Deaerating
open
feedwater
heater
Feed
pump
Closed fw
heater
Feed
pump
2.14 Indicative configuration for a large-scale steam turbine power
plant.
© Woodhead Publishing Limited, 2012
Fundamental principles of concentrating solar power (CSP) systems 43
At a more pragmatic level, managing the chemical composition of the
cycle water is an important part of the process. A fraction of the water is
periodically ‘blown down’ (expelled from the system) to keep the level of
impurities, such as dissolved salts, within acceptable limits. An open feedwater heater involves the direct mixing of the bled steam and feedwater. It
is operated at atmospheric pressure and the heating also has the effect of
driving off dissolved air prior to sending feedwater to the boiler.
Systems are more efficient if they are built as larger units and run at full
load. Most, but not all, of the size efficiency advantage is achieved at the 50
to 100 MWe scale. Larger systems are less costly per unit of capacity. Largescale power-generating turbines used in coal power stations are typically
around 500 MWe. For a CSP application, a larger turbine requires a large
field, which results in extended thermal line losses, and so there is a tradeoff against turbine size, with a 250 MWe unit being suggested by many
observers as offering the lowest cost of energy. As of early 2012, no CSP
systems have been built to this size, although there are several in planning
stages.
The most efficient state-of-the-art steam turbines work at up to 700°C
steam inlet temperature. Trough and linear Fresnel concentrators are,
however, limited to around 400°C if thermal oil heat transfer fluid is used,
and up to 500°C if an alternative HTF such as direct steam generation
(DSG) is used. Tower and dish systems are able to reach the temperatures
needed for the highest possible steam turbine inlet temperatures and pressures; the limitation in that case becomes survival of materials either in the
turbine or in the solar receiver.
State-of-the-art steam turbines are now produced that work at supercritical conditions, for maximum conversion efficiency. Supercritical steam is
steam at pressures and temperatures above the critical point (22 MPa,
374°C); at these conditions, the phase-change occurs continuously rather
than via nucleate boiling, at higher temperatures. Viable only at very large
scales, these turbines have not yet been applied to CSP plants.
A major area of difference between solar and fossil operation of steam
turbines is the intermittent and changing nature of solar input. This has two
potential impacts:
•
•
the wish to vary turbine speeds up and down more frequently and more
rapidly than in steady-state fossil-fuelled operations, and
the wish to run at part-load more frequently.
Whilst inclusion of energy storage can mitigate these effects to some extent,
directly transferring technologies and practices from conventional generation does not necessarily give optimal results. Turbine manufacturers are
now producing steam turbines customized for CSP application, with these
issues in mind. Such steam turbines are able to reach full power within 30
© Woodhead Publishing Limited, 2012
44
Concentrating solar power technology
minutes from a cold start and less for a warm start. Typical steam turbine
heat to AC electricity conversion efficiencies for existing CST plants are
around 40% gross at full load.
Reciprocating steam engines are more efficient than turbines at very
small scales. They are still produced commercially but remain relatively
unpopular due to complexity and maintenance issues.
2.7.2
Organic Rankine cycles
An organic Rankine cycle (ORC), is fundamentally the same as a steam
Rankine cycle; however, it uses a lower boiling point organic fluid to better
match its operation to lower temperature heat sources. ORC systems can
achieve better efficiencies than steam turbines for smaller systems (less than
a few MWe). However, the capital and operating and maintenance (O&M)
costs are higher per installed MW than for a water/steam system. ORC
technology is being actively pursued for geothermal power applications
because of its better match to lower temperature sources. ORC systems
have been applied to a few modest sized linear concentrator CSP systems.
Another potential application for ORC systems in CSP is as a ‘bottoming
cycle’, whereby a high temperature cycle (see discussion of Brayton cycle
below), produces exhaust heat (that would otherwise be wasted) which is
at sufficiently high temperature to drive an ORC system for additional
power generation.
2.7.3
Stirling engines
A Stirling engine is an externally heated engine with reciprocating pistons
that operate on a fixed, enclosed amount of gaseous working fluid, usually
hydrogen or helium or possibly air. The ideal cycle is made up of a mix of
constant temperature and adiabatic (zero heat transfer) processes. In the
ideal limit, it is capable of the highest thermodynamically possible conversion efficiencies between two constant temperature limits.
The Stirling engines contemplated for CST applications to date have
all been small (in the tens of kWe range), although large, fossil-fuelled
systems for marine propulsion do exist. Dish-mounted Stirling engines
incorporate receiver, engine and generator in a single package at the focus
(see Chapter 9).
Stirling engines have long been applied to dish concentrators. This long
history and predominance leads many in the CSP field to refer to dish
systems in general as ‘dish-Stirling’, even though dishes have been applied
to direct steam generation, photovoltaics and other approaches. Stirling
engines have not so far been applied in any serious way to other collector
types.
© Woodhead Publishing Limited, 2012
Fundamental principles of concentrating solar power (CSP) systems 45
There are two types of Stirling engines: piston-crankshaft types built in
a similar manner to internal combustion engines and free-piston engines
which involve an oscillating piston attached to a linear generator, but with
no physical restraining linkage. Both types have been used in practice.
Dish-Stirling systems continue to hold the record for the highest solar to
electric conversion efficiency of any technology, with total solar to AC
electric efficiencies of around 30% at design point DNI, reliably reproduced. Stirling systems can be used for much smaller systems than Rankine
cycles, but in the dish-Stirling configuration, thermal storage is yet to be
developed.
2.7.4
Brayton cycles
The Brayton cycle is the basis of jet engines and the turbo generators used
in gas turbine power stations. It is a common misconception that ‘gas turbines’ are so named because they burn gas; however, the name actually
refers to the fact that a gas (usually air) is the working fluid. In fuel-fired
mode, any hydrocarbon fuel, such as aviation fuel, diesel, LPG, propane or
bio gas, could be burnt to achieve the required heating. Alternatively solar
heat could be used to raise the temperature of the compressed air before
expansion. With temperatures before expansion of around 1,000°C needed
for efficient operation, this is only likely to be achieved with tower systems
or dishes. Demonstration CSP systems using solar heating of a Brayton
cycle have been operated.
In fossil-driven applications, a combined cycle power plant uses a gas
turbine with its high temperature exhaust gases then directed to a ‘heat
recovery steam generator’ that provides steam for a steam turbine cycle.
Potentially the combined conversion efficiency is in excess of 50% and
represents the highest thermal to electric conversion efficiency solution
currently available commercially. A major attraction with applying the
Brayton cycle to CSP applications is to also implement combined cycle
operation with either steam or ORC bottoming cycles in a similar high
efficiency manner.
For dish applications, the Brayton cycle offers the potential of reduced
O&M costs compared to Stirling engine systems. An area of current solar
thermal energy research interest is in supercritical carbon dioxide Brayton
cycles (s-CO2 cycles). The different thermodynamic properties of CO2 compared to air allow higher overall cycle efficiencies to be achieved.
2.7.5
Concentrating photovoltaics
Concentrating photovoltaic (CPV) systems are discussed in detail in
Chapter 10. They typically use expensive, high efficiency cells to gain
© Woodhead Publishing Limited, 2012
46
Concentrating solar power technology
maximum advantage from the investment in concentrator systems. Highquality, single-crystal silicon cells with efficiencies of around 20% have been
utilized. Going beyond this, cells, such as the multi-junction cells developed
for space applications, have had a rapid efficiency increase over the last
decade (from 30% efficient to 43% efficient).
With a CPV system, there are parasitic losses relating to tracking system
operation, controls, wiring losses, inverter efficiencies and operation of the
cooling system. These parasitic losses reduce the useful AC output of the
entire system.
A key issue with any high concentration PV system is the heat that results
from the photons that are not converted directly to electricity. At 500 suns,
a triple-junction cell would be destroyed within a few seconds without a
highly efficient cooling system to extract the heat.
2.7.6
Others
There are other ways of converting solar radiation to electricity that may
eventually be competitive, for example:
•
•
•
•
•
The Kalina cycle is a modified Rankine cycle involving mixtures of
ammonia and water of varying concentration, that offers higher performance for temperatures between 200 and 300°C. It is being pursued
commercially and also targeted at geothermal power applications.
Thermoelectric converters produce electricity directly from heat. Semiconductor-based systems work in an analogous way to photovoltaic
cells, except that excitation of electrons into the conduction band is
via thermal excitation rather than individual photon absorption.
Thermionic converters also produce electricity directly from heat; they
excite electrons from an active surface across an evacuated region to a
collector.
Thermo-photovoltaics use PV cells tailored to thermal radiation wavelengths to convert the radiation re-emitted from heated surfaces.
Magneto-hydrodynamic converters use the expansion of heated ionized
gas through a magnetic field to generate a potential difference.
Whilst this book is primarily directed at electric power generation, final
conversion can also be to drive a chemical process such as discussed in
Chapter 20 or for process heat as discussed in Chapter 19.
2.8
Maximizing system efficiency
Previous sections have discussed optical errors, concentration ratio, and
the sources of thermal losses from a receiver. Performance of these
© Woodhead Publishing Limited, 2012
Fundamental principles of concentrating solar power (CSP) systems 47
subsystems are combined, into an overall system efficiency (Eq. [2.1],
repeated here):
ηsystem = ηoptical × ηreceiver × ηtransport × ηstorage × ηconversion.
A common convention is for the concentrator/optical efficiency ηoptical to
include all reflectivity and flux spillage losses up to the point where radiation enters the receiver. The receiver efficiency ηreceiver includes both reflective losses from the receiver as well as radiative, convective and conductive
losses. Transport efficiency ηtransport includes thermal losses from pipework
conveying the heat transfer fluid to either the storage or the power cycle
(or the chemical or industrial heating process), as well as losses in heat
exchangers between the HTF and the power cycle working fluid. Storage
efficiency ηstorage, as an energy efficiency, accounts for thermal losses from
the storage vessel as well as possible losses in pressure or chemical potential, depending on the type of storage; it may also include losses in heat
exchangers between the HTF and the storage medium, and between the
storage medium and the power cycle working fluid. Finally, the power cycle
efficiency ηconversion (or conversion efficiency) accounts for the thermal, friction, and electrical losses in the power cycle, noting that this is limited by
the second law of thermodynamics as discussed below.
It must be noted that the overall system efficiency is not maximized by
individually maximizing each of the component efficiencies; the maximum
is found only by examining all effects together.
2.8.1
The second law of thermodynamics and
exergy analysis
Maximizing overall system efficiency in energy terms is an appropriate goal.
However, consideration of the issues in purely energy terms does not
provide all the information needed for optimization.
The second law of thermodynamics addresses the inherent of irreversibility of processes and leads to the ultimate limiting conversion efficiency
of a heat engine between isothermal reservoirs, given by the well-known
Carnot limit on the efficiency, η
η≡
W QH − QL
Q
T
=
= 1− L = 1− L .
QH
QH
QH
TH
[2.45]
In this equation, temperatures are absolute temperatures, expressed in SI
units of kelvin, with 0 K = −273.15°C. Clearly operating at low temperatures
to achieve high receiver efficiencies does not result in efficient power
generation.
© Woodhead Publishing Limited, 2012
48
Concentrating solar power technology
An ideal Stirling engine is the only power cycle that has a conversion
efficiency as indicated by the Carnot expression. In a receiver-mounted
application at steady state, a Stirling engine operates between a fixed
high temperature receiver and a fixed low temperature environment.
However, for systems where a HTF is employed between receiver and
power cycle, the power cycle is no longer operating between effective
isothermal reservoirs. The Carnot efficiency quantifies the special case of
the maximum amount of work that can be extracted between two isothermal reservoirs. It can be used further to develop a very useful thermodynamic concept called ‘exergy’.18 Textbooks such as Çengel and Boles
(2010) or Moran and Shapiro (2011) give essential background theory for
this area.
The exergy Φ of a system is defined as the maximum amount of work
that could usefully be done whilst bringing the (thermodynamic) system to
equilibrium with its surroundings. In this context, a system could simply be
a particular unit mass of a substance such as a heat transfer fluid. Following
on from this definition, every energy flow can be associated in a quantitative
sense with a corresponding exergy flow, being the ideal amount of work
that could be extracted from the energy flow. Exergy analysis is a powerful
tool in understanding and improving CSP systems which are designed to
produce power as their ultimate output.
Each subsystem efficiency in Eq. [2.1] can have a corresponding
exergetic efficiency defined. Since Eq. [2.1] as it stands includes a final
output form of electricity (pure exergy), maximization of either overall
energy or exergy efficiency will occur for the same design and operating
parameter values.
Table 2.1 illustrates the temperature dependence of the Carnot efficiency
together with the exergy-to-energy ratio of some potential HTFs at the
same temperature. It can be seen that operating thermal systems at realistically achievable temperatures represents a major step down in the potential
to extract work compared to the sun’s surface temperature. It can also be
seen that transfer of heat from an isothermal source to a real HTF heated
to the same temperature also reduces the potential to extract work. That is,
it is an irreversible process that destroys exergy.
It can be seen that, for a given temperature, the Carnot efficiency for
operation from an infinite isothermal source is higher than the exergy-toenergy ratio of the alternative working fluids. Energy in a finite amount
of a substance at a particular temperature can be thought of as being
available over the range from its temperature down to ambient temperature
18
Exergy is also sometimes referred to as availability. This should not be confused with availability as a term used in the power industry for the fraction of time that a piece of equipment
is available to function on demand.
© Woodhead Publishing Limited, 2012
© Woodhead Publishing Limited, 2012
5,527
1,000
700
400
300
200
Temperature
(°C)
5,800
1,273
973
673
573
473
Temperature
(K)
94.9
77.0
69.9
56.5
48.9
38.0
Carnot
efficiency
(%)
84.1
56.1
48.3
35.9
29.8
22.0
Exergy-energy
ratio of a
constant specific
heat HTF (%)
48.0
43.4
31.5
Exergy-energy
ratio of steam
at 15 MPa (%)
Table 2.1 Thermodynamic efficiency metrics relative to an environment of 298 K, 1 bar
44.6
40.2
39.0
Exergy-energy
ratio of steam
at 5 MPa (%)
n.a.
n.a.
40–45
35–40
25–30
Typical real
steam cycle
net efficiency
(%)
50
Concentrating solar power technology
according to the amount needed for each increment of temperature rise.
Steam has a profile that is strongly affected by the phase change from water
to steam. Depending on how close this is to the actual steam temperature,
its exergy-to-energy ratio is either increased or decreased relative to a
constant specific heat fluid.
The concept of regenerative feedwater heating in a steam cycle using bled
steam is best understood from an exergy point of view. Partially expanded
steam has already given up some of its exergy to power generation, so using
it to pre-heat feedwater destroys less exergy than allowing the highest
temperature heat source to heat the feedwater from ambient temperature.
Applying this to a CSP system, then, we can say that sending the HTF to
the receivers at a higher temperature increases the potential to generate
work from the energy that is absorbed in the receivers. Doing so, however,
means receiver average temperature and associated losses are increased
and higher HTF flow rates must be maintained.
2.8.2
Heat exchange between fluids
A major cause of exergy loss in CSP systems is any place in a system where
heat transfer from one fluid to another occurs, usually in a heat exchanger.
Major areas of heat exchange in CSP systems can include:
•
•
•
•
from HTF to storage medium
from HTF to power cycle working fluid
from storage medium to power cycle working fluid
from power cycle working fluid to ambient air (condenser or cooling
tower).
Well-designed heat exchangers create a large interface area between two
fluid streams, such that heat transfer can be achieved with a lower temperature drop, even if the overall heat transfer coefficient is not high. To build
such a heat exchanger can be extremely expensive however, so a costbenefit optimization is usually required.
Systems that avoid unnecessary heat exchange steps are receiving
increased attention, including in areas of direct steam generation, and
molten-salt parabolic trough fields.
Irrespective of the overall heat transfer coefficient achieved in a heat
exchanger, the fluids involved may not have the same profiles of enthalpy
increase to temperature increase. This is particularly true if one has a phase
change (water to steam). This can be a major source of exergy loss.
In complex processes with multiple heating and cooling requirements, the
technique of pinch analysis, attempts to combine all heat transfer temperature profiles in an optimal manner.
© Woodhead Publishing Limited, 2012
Fundamental principles of concentrating solar power (CSP) systems 51
0.8
C = 20
C = 40
C = 400
C = 1,000
C = 2,500
C = 10,000
Receiver efficiency hrecv
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0.0
500
1000
1500
2000
2500
Receiver temperature Tr (K)
3000
2.15 Efficiency of a simplified solar collector with 100% optical
efficiency, average absorptivity of 0.8 (assumed equal to emissivity)
plotted both with no external convective loss (solid lines) and with
20 W/m2K external convective loss (dotted lines). DNI is 800 W/m2,
ambient temperature 300 K, and sky temperature 270 K.
2.8.3
Optimization of operating temperature
A simplified analysis of the performance of a solar thermal energy system
is useful in demonstrating the competing design demands. From the theory
of reversible heat engines, we know that power cycles running from hightemperature thermal reservoirs are most efficient, but we also know that
thermal losses increase as receiver temperatures increase.
Using the simplified expressions for the various loss mechanisms, and
some typical parameter values, gives the results for receiver efficiency as a
function of concentration ratio and temperature shown in Fig. 2.15. It can
be seen that at low temperatures, all efficiencies reduce to the average
absorptivity value. As temperature is increased,19 efficiencies drop with
higher concentration systems having higher receiver efficiencies at any
given temperature. The various curves intersect the x-axis at the corresponding stagnation temperature.
19
In a real system, operating temperature can be controlled through variation of HTF flow
rate, for example.
© Woodhead Publishing Limited, 2012
52
Concentrating solar power technology
To simulate an idealized operating regime for maximizing power production, the receiver efficiency must be multiplied by a power cycle conversion
efficiency to yield an overall ideal system efficiency.20 To a large extent,
power cycle efficiency dependence on operating conditions is complex and
must be established from empirical performance curves. The difference in
exergy value of the HTF before and after transfer of heat to the working
fluid at input to the power cycle provides a good indicator, with real power
cycles extracting something over 70% of the available exergy as work. The
Carnot expression provides a good qualitative illustration of the principles,
however.
If the receiver efficiency results of Fig. 2.15 are multiplied by the Carnot
efficiency pertaining to the receiver temperature at each point to give a
semi-ideal system efficiency, the results are as shown in Fig. 2.16. It can be
seen that there is now a clear peak efficiency at a particular temperature
for each concentration ratio. Higher concentration ratios have higher peak
efficiencies and these occur at higher receiver temperatures.
0.8
C = 20
C = 40
C = 400
C = 1,000
C = 2,500
C = 10,000
hcarnot
System efficiency hsys
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0.0
500
1000
1500
2000
2500
Receiver temperature Tr (K)
3000
2.16 Ideal system efficiency as a function of receiver temperature.
Carnot efficiency is also shown as a dashed line as a function of
receiver temperature. The dotted lines include the effect of convective
heat loss, for a fixed convection coefficient as described in Fig. 2.15.
20
This is equivalent to treating optical efficiency and energy transport and/or storage efficiency
as 100%.
© Woodhead Publishing Limited, 2012
Fundamental principles of concentrating solar power (CSP) systems 53
2.8.4
Optimization of aperture size
Receiver aperture size is an important parameter to optimize. Radiative
and convective losses are essentially proportional to aperture area, for
receiver temperature held constant, suggesting that this area should be
minimized. On the other hand, a larger aperture will intercept more of the
incident radiation so increasing the optical efficiency. For a given focalregion flux distribution and operating temperature, there will be an aperture size beyond which any further increases will increase thermal losses
by more than the collected energy is increased.
This choice of optimum aperture and operating temperature will be different for every concentrator and application. To illustrate the idea, the
following results are taken from Steinfeld and Schubnell (1993). They analysed the optimum choices based on an actual flux distribution measured
with a small solar furnace. Figure 2.17 shows the actual focal plane flux map
as a contour plot. Steinfeld and Schubnell assumed only radiation losses
from a hypothetical receiver and treated it as a uniform temperature blackbody disc. On this basis they defined an absorption efficiency equal to the
product of concentrator and receiver efficiency. They determined this for a
15
50
25
10
0
425
3 40
30 50 0
0
0
25
0
20
150
5
15
0
50
50
0
45
100
75
Vertical distance (cm)
30
0
350
25
0
10
150
0
20
0
25
10
75
0
10
75
75
25
50
25
(W/cm2)
0
0
5
10
Horizontal distance (cm)
15
2.17 Solar flux distribution measured at the focus of the Paul Scherer
Institute solar furnace in October 1990. The power intercepted by the
aperture can be found by integrating solar flux through the circled
area (reproduced from Steinfeld and Schubnell, 1993).
© Woodhead Publishing Limited, 2012
54
Concentrating solar power technology
Absorption efficiency (%)
100
300 K
1200 K
1500 K
80
1800 K
60
2100 K
40
2400 K
20
2700 K
0
1
2
3
4
5
6
Aperture radius (cm)
7
8
2.18 Energy absorption efficiency as a function of aperture radius for
various receiver temperatures. The dotted line is the locus of
maximum efficiency, for which optimum aperture radii are determined
(reproduced from Steinfeld and Schubnell, 1993).
range of receiver temperatures and radii and produced the results shown
in Fig. 2.18. For each operating temperature, the absorption efficiency first
increases as aperture size is increased and more radiation is intercepted,
but it then reaches a peak and decreases once the increased radiation loss
from area increase outweighs the extra incident radiation intercepted. For
any particular operating temperature, there is an optimum aperture radius.
Higher temperatures with associated higher levels of re-radiation dictate
smaller aperture size. Choice of higher operating temperature also limits
the maximum absorption efficiency that can be achieved.
Taking this further to a semi-ideal overall system efficiency by multiplying the absorption efficiency by the Carnot efficiency at the receiver temperature yielded the results in Fig. 2.19 and shows that an ideal Carnot cycle
driven by a black-body receiver on this collector would have a peak overall
efficiency when operated at 1,200 K with a 7 cm radius aperture.
2.8.5
Solar multiple and capacity factor
It is the overall average system efficiency which is most important, rather
than the design point steady state efficiency. The concepts of capacity factor
and solar multiple are important in this respect. The capacity factor, CF, of
a piece of equipment is its fractional utilization calculated over a long time
period. In the case of a CSP system, full utilization corresponds to the power
© Woodhead Publishing Limited, 2012
Fundamental principles of concentrating solar power (CSP) systems 55
3000
hoverall (%)
2000
30
Temperature (K)
10
2500
10
20
40
50
1500
60
1000
65
50
40
500
60
20
0
0
2
4
6
Aperture radius (cm)
8
10
2.19 Overall system efficiency as a function of the aperture radius and
temperature, using the solar flux distribution of Fig. 2.17 (reproduced
from Steinfeld and Schubnell, 1993).
block running at. its nominal (rated) design point or ‘nameplate’ electrical
output power Wdes,pb. A capacity factor of 100% means that the plant is
running at full rate all of the time; 25% capacity factor means
that the long.
term average of the power block electrical output Wavg,pb is 25% of the
design point output
CF =
avg
W
.
Wdes, pb
[2.46]
For CSP systems without energy storage or backup boilers, the turbine will
not be operating at night, nor during periods of cloudy weather; for systems
such as this, the capacity factor will be approximately 20%, purely due to
sun position and weather effects. With the addition of energy storage,
the capacity factor can be increased greatly, up to as much as 75% or
more, allowing round-the-clock operation for much of the sunnier part of
the year.
The capacity factor for the power block of a CSP system can also be
increased by oversizing the solar field (more heliostats or troughs per steam
turbine, for example). This leads to the definition of the solar multiple, SM,
which
is the ratio of the solar field design point thermal power output
.
Qdes,field (normally calculated at solar noon on a clear midsummer day) to
© Woodhead Publishing Limited, 2012
56
Concentrating solar power technology
the thermal
. power demand of the power block when running at its nominal
capacity Qdes,pb.
SM =
Q des, field
.
Q des, pb
[2.47]
When a thorough economic analysis (Section 2.10) is performed on a CSP
plant, it will typically be found that an optimal solar multiple for a system
without storage will be in the range 1.15 to 1.30, while for systems with
storage, the optimal solar multiple will be 2 or higher, depending on how
much storage is installed. This is because there are efficiency and capital
cost penalties associated with a power block that frequently runs at partload; it is better to waste a little of the collected heat during the maximal
solar periods in exchange for running the power block at its design point
capacity, at consequently higher efficiency, for more of the year.
2.9
Predicting overall system performance
Predicting the output of a CSP system with reasonable accuracy is a complex
process. Thermal systems include multiple subsystem components with
thermal capacity whose behaviour at any point in time depends not only
on the instantaneous conditions the whole system experiences, but also the
recent history of its operation. Power cycles have efficiencies that change
with load and can take significant time to start up and shut down.
There is a range of approaches to modelling CSP systems and it is an
ongoing area of R&D. One can distinguish between those that model on
half to one hour time steps and treat most thermal components as being in
‘pseudo steady-state’ and more fundamental approaches that attempt to
track short duration cloud and thermo-fluid transients.
A recent review of CSP system simulation work, together with an overview of background theory for simulation of a complete plant including
storage is given by Llorente García et al. (2011). Standardization and benchmarking of CSP simulation tools is currently in progress by the SolarPACES
organization (Eck et al., 2011). Commercial software for energy system
modelling which has been used in solar thermal applications includes
IPSEpro, Ebsilon, EcoSimPro, TRNSYS, GateCycle, Dymola, Mathematica
and Aspen. The System Advisor Model (SAM) from NREL (NREL, 2012)
is an easy-to-use, free, but closed-source package that uses the well-known
TRNSYS simulation engine internally. Possibly the only complete free
open-source system model is SOLERGY (Stoddard et al., 1987). Key
outputs from such analyses include annual power output, capacity factor.
To obtain optimal sizing and operational strategies, the analysis must be
coupled with an economic analysis (Section 2.10).
© Woodhead Publishing Limited, 2012
Fundamental principles of concentrating solar power (CSP) systems 57
2.9.1
Case study using the system advisor model (SAM)
As an example of a system modelling analysis, results for a trough system
modelled with the System Advisor Model (SAM) are presented here. SAM
can predict hourly, monthly and annual output of CST, CPV, flat PV and
also a range of other renewable energy systems. There has been an extensive
body of work around its application to CSP systems in particular.
The key inputs for system performance forecasting are the direct normal
irradiation (DNI) time series data, together with the associated ambient
temperature, humidity and wind speeds. SAM accepts weather files in a
range of ‘typical meteorological year’ (TMY) formats, meaning an artificial
year assembled from real months from real years that yield a match with
overall long-term averages, as well as containing a representative range of
unusual/extreme days. Generally the expectation is that the TMY file will
provide hourly data.
There is a range of available templates and files of predetermined case
studies for use with SAM. One of these is a verified model of the actual
Nevada Solar 1 64 MWe trough system that is located near Las Vegas. The
key parameters of the plant are:
• 64 MWe nominal electrical power output
• no storage
• solar multiple of 1.264 (i.e. solar field is oversized relative to power block
system capacity at design conditions)
• total trough aperture (collector) area of 357,428 m2.
Figure 2.20 shows the daily profiles of DNI and net generation levels for a
clear day and two days that show partial cloud events in the afternoon.
Several key effects can be seen:
•
•
•
•
Prior to start-up each day, there is a period of negative generation (net
energy consumption) as the system is prepared for operation.
During the morning start-up phase, generation levels lag behind the
rising DNI levels, reflecting the inertia of the system.
During periods of continuous high DNI levels, system output can vary,
presumably reflecting the variation in factors like ambient temperature
affecting overall performance.
For DNI levels below a threshold level, no generation is produced.
Figure 2.21 shows the results for predicted monthly output for this system
at its Las Vegas, USA, site and three alternative sites in Australia. Key
parameters and outputs for these cases are listed in Table 2.2.
The northern hemisphere site has output peaking in mid-year whereas
the southern hemisphere Australian sites have output peaking at the beginning/end of the calendar year. The three Australian sites span from northern
© Woodhead Publishing Limited, 2012
58
Concentrating solar power technology
1,000
80
DNI
800
DNI (W/m2)
600
40
400
20
200
0
Net generation (MW)
60
Net
generation
–20
0
April 28
April 29
April 30
2.20 Three days of modelled output of a 64 MWe parabolic trough
system using Longreach, Australia, weather data points.
MWh/month
64 MW Trough plant
18000
16000
14000
12000
10000
8000
6000
4000
2000
0
Longreach
Mildura
Moree
Las Vegas
1
2
3
4
5
6
7
Month
8
9
10
11
12
2.21 Modelled output of a 64 MWe parabolic trough system sited at
Las Vegas (USA) and Longreach, Moree and Mildura (Australia).
to southern latitudes. It is apparent that the Longreach site, which is closest
to the equator,21 shows the most uniform output through the year.
Overall annual generation correlates with annual DNI, but in a nonlinear manner. Referring to Table 2.2, it can be seen that Longreach outperforms the Las Vegas site, even though it has slightly lower DNI.
Conversely, Mildura, with around 20% lower DNI than the best sites, shows
output reduced by around 40%. High ambient temperatures and high
21
Whilst being closer to the equator improves the performance in inland Australia, note that
equatorial regions generally offer poor performance due to tropical cloud and humidity.
© Woodhead Publishing Limited, 2012
© Woodhead Publishing Limited, 2012
117,147
128,794
106,165
89,714
Longreach
(North Queensland,
Australia)
Moree
(NSW, Australia)
Mildura
(North Victoria,
Australia)
Net annual
generation
(MWhe)
Las Vegas
(Nevada, USA)
Location
11.81%
13.17%
14.05%
12.57%
Annual
conversion
efficiency
16.00%
18.90%
22.90%
20.90%
Capacity
factor
2,124.6
2,254.6
2,564.4
2,606.6
DNI
(kWh/m2/yr)
−115.167
144.28
149.83
142.08
−23.43
−29.48
−34.23
Longitude
36.08
Latitude
16.9
18.7
23.9
19.5
Average
ambient
temp. (°C)
Table 2.2 Modelled outputs of a 64 MWe parabolic trough system in Las Vegas and at various sites in Australia
3.4
2.3
2.4
4.1
Average
wind speed
(m/s)
60
Concentrating solar power technology
average wind speeds would work to reduce the system output; however, the
variation in these parameters does not appear significant between the sites.
Another differentiator is the extent to which low DNI days are made up of
short intervals of broken cloud or whole days of no sun. Broken cloud
works against output for CST systems, since the time taken for the system
to reach operating temperatures makes operation extremely inefficient in
such circumstances.
2.10
Economic analysis
The discussions above have suggested that CSP system design must consider cost trade-off issues as well as simply maximizing efficiencies. This
section introduces the relevant aspects to complete the picture. A definitive
description of methodologies for the financial analysis of energy systems is
given by Short et al. (1995).
CSP systems have high capital costs and no fuel costs. Initial investments
must be amortized over the working life of systems. Thus the key issue is
the net present value (NPV) over the lifetime. The basic formula for evaluating NPV is
N
Ci
i
i = 1 (1 + DR)
NPV = ∑
[2.48]
where the cash flows Ci are those occurring in time interval (year) i, DR is
the discount rate and N is the total number of compounding periods.22
Cash flows can be measured in either nominal or real currency units. A
real cash flow is adjusted for inflation and expressed in a currency value of
a specific year irrespective of the year it takes place and thus has a constant
effective value. The discount rate can be either nominal or real. NPVs can
be calculated using real currency cash flow measurements together with real
discount rates, or nominal currency cash flow measurements with nominal
discount rates; the same NPV will be obtained in either case.
For a CSP system, the key cash flows are the initial capital investments
(negative), ongoing O&M costs (negative), the costs of ongoing inputs such
as fuel for hybrid operation or water for cooling (negative), revenue from
direct energy sales (positive) and possible provision of ancillary services
(positive). Key parameters are the discount rate and the assumed lifetime
of plants, both of which have a significant impact on overall NPV results. A
longer assumed plant life and a lower discount rate both work to increase
22
This is the most commonly recognized form of NPV on the assumption of annual compounding. Compounding can actually be done on any time scale including continuously. Also, in a
strict mathematical sense, DR is a fraction per unit time and is multiplied by the compounding
time interval (in this case 1 year).
© Woodhead Publishing Limited, 2012
Fundamental principles of concentrating solar power (CSP) systems 61
NPVs for renewable generation. If the ‘marketplace’ assesses that a project
or technology is ‘high risk’ this leads to the use of shorter lifetimes for
amortization and application of higher discount rates.
The levelized cost of energy (LCOE, also known as levelized energy cost,
LEC) is the most frequently used economic performance metric for power
generation plant. It is a standard metric used not just for CSP or other
renewable energy systems but for any form of generation technology. It is
defined as the constant per unit cost of energy which over the system’s
lifetime will result in a total NPV of zero. In other words it is the ‘break
even’ constant sale price of energy. Thus
(
N
NPV = ∑ 1 (LCOE × Eannual × (1 − T ))
(1 + DR) j
)
− NPVLCC = 0,
[2.49]
where T is the tax rate and NPVLCC is the net present value of all lifecycle
costs and Eannual the annual generated electrical energy. This gives
LCOE =
∑
N
1
(
NPVLCC
(Eannual × (1 − T ))
(1 + DR)
j
)
.
[2.50]
From a purely societal perspective, tax would be omitted from the calculation; however, from a plant owner/business perspective, it is assumed that
tax would be applied to energy sales, whilst various tax deductions would
work to reduce costs.
For fossil fuel power systems, a major part of the lifecycle costs will be
the fuel that is consumed on an ongoing basis in inverse proportion to the
conversion efficiency. For CSP and other renewable systems, the dominant
lifecycle cost is the initial capital investment, with ongoing operations and
maintenance costs as a small but significant contributor.
LCOEs can be in real (inflation independent) or nominal terms, which
can be confusing because they are expressed in year 0 dollar values in either
case. A nominal LCOE represents a hypothetical income that declines in
real value year by year, whereas a real LCOE has a constant ‘value’. Since
the total NPV using either method must be the same by definition, the
nominal LCOE will be the higher of the two. Real LCOEs are typically
used for future long-term technology projections, whereas nominal ones are
often used for short-term actual projects.
From a pure societal perspective, it can be argued that tax issues can be
left out of the LCOE. However, from the perspective of a commercial entity
owning a system, the prevailing assumption is that, in order to break even,
it must be assumed that energy produced is taxed at the standard corporate
tax rate. Against this, interest, depreciation and operating costs are tax
deductible.
© Woodhead Publishing Limited, 2012
62
Concentrating solar power technology
Detailed, project-specific LCOE evaluations are based on complex calculations summing every discounted cash flow over the system lifetime,
which are then solved iteratively to establish the real dollar value of energy
which gives the total NPV of zero. Issues that are typically encountered
include:
•
•
•
•
•
•
•
•
debt financing (loans) may be paid off over a different timescale to
equity investments
tax benefits may apply in different jurisdictions
tax deductible depreciation may apply over a shorter timescale than the
project
construction is staged over several years and subject to higher interest
rates for finance
system output may take some time to stabilize as commissioning processes proceed after first start-up
system output may be subject to other predictable variations over time
(such as a component with known degradation rate)
major plant upgrade expenditures may be predicted at certain times in
addition to overall continuous O&M
various inputs may be subject to different escalation rates.
All these issues are project-specific, depending on technology type, developer status and site location.
Studies that report the LCOE for CSP systems and other generation
types are often poor at documenting all of their input parameter assumptions and describing the methods used in a comprehensive way. In many
cases, the methodology is actually intentionally withheld as it is embodied
in proprietary financial models.
A methodology that is somewhat simplified but has sufficient complexity
to allow issues of tax, cost of equity and cost of debt to be examined, is
based on a life cycle costs NPV calculation embodied in the following
formula
NPVLCC = EQ − ∑ 1
ND
DEP × T
+ ∑1
NL
LP
(1 + DR)
(1 + DR )
N A O × (1 − T )
SV
+ ∑1
−
,
(1 + DR ) j
(1 + DR )N
j
− ∑1
NL
j
INT × T
(1 + DR ) j
[2.51]
where EQ is the initial equity contribution from the project developer, DR
is the nominal discount rate, ND the period (number of years) over which
the system can be depreciated for tax purposes, DEP is the amount of
depreciation in a year, T is the taxation rate applying to assumed income
from energy sales, LP is the annual loan payment, INT is the reducing
amount of interest paid each year as the loan is paid off, NL is the term
(number of years) of the loan, AO is the annual operations cost which could
© Woodhead Publishing Limited, 2012
Fundamental principles of concentrating solar power (CSP) systems 63
be calculated from fixed and variable maintenance contributions and any
fuel costs as needed, N is the project lifetime23 and SV is the end of project
life salvage value.
The simplifying assumptions used are as follows:
•
•
•
•
•
•
The analysis begins from the time of plant commissioning.
Annual energy production is assumed constant over project life.
The equity contribution is assessed at the beginning of year 1 and so is
assumed to have all costs of construction finance rolled into it.
Depreciation is linear in nominal dollars.
Loan payments are constant for each year of the loan and are in nominal
dollars based on amortization of a debt across a loan term.
Annual O&M costs are constant per year in nominal dollar terms across
project life. (This is possibly the most significant, since it does not reflect
the lumpy expenditure likely on component overhaul).
To aid in understanding, LCOE can be simplified further if tax is not
considered and the cost of capital (both debt and equity) can be rolled into
a single discount rate and the debt and equity investments rolled into a
single capital cost. If fixed and variable operation costs are separated out,
the result is
LCOE =
(FR + O&M fixed )C0
C fuel
+
+ O&M variable
ηconversion
PFc
[2.52]
where P is the nominal design point electrical power capacity of the system,
Fc is the capacity factor (the annual average fraction of nominal capacity
achievable), O&M are operation and maintenance costs that are split
between those that are in proportion to generation (variable, expressed in
the same units as the LCOE) and those that are fixed annual costs (expressed
as a fraction of capital cost per year), Co is the total initial capital cost, Cfuel
is the per unit energy cost of any fuel used in a hybrid system, hconversion is
the conversion efficiency of fuel to electricity and
⎛ DR(1 + DR)n ⎞
FR ≡ ⎜
⎝ (1 + DR)n − 1 ⎟⎠
is the capital recovery factor.
The capital recovery factor (sometimes called annualization factor) is
dimensionally the same as the discount rate and represents a rate of repayment that covers return on investment (at the assumed discount rate) plus
paying off the capital in the system’s lifetime. The dependence of capital
recovery factor on discount rate and system life is illustrated in Table 2.3.
23
In a literal sense, a plant may be decommissioned at the end of its useful life. However, if a
conservative assumption has been made, it may prove to be the case that the plant’s life actually exceeds the value assumed.
© Woodhead Publishing Limited, 2012
64
Concentrating solar power technology
Table 2.3 The dependence of capital recovery factor on discount rate for
system lifetimes of 20, 25 and 30 years
Discount rate
(%/year)
Capital recovery
factor for
20-year life
(%/year)
Capital recovery
factor for 25-year
life (%/year)
Capital recovery
factor for 30-year
life (%/year)
3.00
4.00
5.00
6.00
7.00
8.00
9.00
10.00
11.00
12.00
6.72
7.36
8.02
8.72
9.44
10.19
10.95
11.75
12.56
13.39
5.74
6.40
7.10
7.82
8.58
9.37
10.18
11.02
11.87
12.75
5.10
5.78
6.51
7.26
8.06
8.88
9.73
10.61
11.50
12.41
The correspondence between installed capital cost, discount rate and LCOE
is shown in Table 2.4 for a 25-year life and O&M costs expressed purely as
a fixed cost of 1% of capital cost per year.
CSP research and development efforts are all essentially directed at
reducing construction costs, improving efficiencies and increasing capacity
factors, these being the technical options for reducing the LCOE.
2.10.1 Stochastic modelling of CSP systems
As a CSP project approaches the point of being financed and constructed,
investors will require analysis of increasing detail and accuracy. In other
words, the confidence interval for the LCOE needs to be known. If done
rigorously, this requires an analysis of the propagation of errors from all
measurement and data sources (including estimated values). These include
errors in weather data, fluid properties including specific heat capacity of
storage media and HTF, power cycle performance and collector and pipework heat transfer coefficients. Recently the SAM simulation tool has
incorporated the ability to perform analysis such as this, and similar capability is offered by some other simulation tools (Ho, 2008).
2.11
Conclusion
This chapter has presented the fundamental principles of CSP systems by
tracing the flow of solar energy from initial collection, through to final conversion to electricity, and has considered the limitations that arise in each of
the subsystems of concentrator, receiver, transport, storage and conversion.
© Woodhead Publishing Limited, 2012
Fundamental principles of concentrating solar power (CSP) systems 65
Table 2.4 The dependence of LCOE on discount rate and capital cost for a
25-year life, 20% capacity factor and fixed O&M costs at 1% of capital cost per
year. The currency units can be interpreted as US or Australian dollars, or
euros, providing that the currency of the LCOE is the same as the currency of
capital cost used
Specific capital cost for
20% capacity factor ($/kWe)
1,000
2,000
Discount rate (%/year)
3.00
4.00
5.00
6.00
7.00
8.00
9.00
10.00
11.00
12.00
LCOE (c$/kWh)
4.6
7.7
5.7
8.4
6.8
9.2
8.0
10.1
9.1
10.9
10.3
11.8
11.4
12.8
12.6
13.7
13.7
14.7
14.8
15.7
3,000
4,000
5,000
6,000
11.5
12.7
13.9
15.1
16.4
17.8
19.1
20.6
22.0
23.5
15.4
16.9
18.5
20.1
21.9
23.7
25.5
27.4
29.4
31.4
19.2
21.1
23.1
25.2
27.3
29.6
31.9
34.3
36.7
39.2
23.1
25.3
27.7
30.2
32.8
35.5
38.3
41.2
44.1
47.1
The basic physical principles are derived from the principles of optics, heat
transfer and thermodynamics. Exergy analysis provides a valuable source of
insight and a tool to further optimize performance during design.
Finally, use of discounted cash flow analysis to derive levelized cost of
energy (LCOE) has been introduced. Ultimately, it is the cost of energy
relative to the income that can be earned that is paramount for CSP systems.
Thus system optimization is a problem that is thermo-economic in nature.
In addition to this, and not covered by this chapter, CSP systems are
informed by a range of other engineering disciplines. Mechanical design,
materials, wind loads, control systems, etc., are all encountered. CSP systems
are a classic example of interdisciplinary systems engineering.
2.12
Sources of further information and advice
In addition to references that have already been cited and which are listed
below in Section 2.13, a number of excellent books have been written in
the fields of solar energy in general, solar thermal energy and also concentrating systems specifically, all of which can offer extra insights. Some of
them are now unfortunately out of print. They include:
Becker M, B Gupta, W Meinecke and M Bohn (1995) Solar Energy
Concentrating Systems, Applications and Technologies, C.F. Mueller
Verlag, Heidelberg.
© Woodhead Publishing Limited, 2012
66
Concentrating solar power technology
Casal F G (1987) Solar Thermal Power Plants, Springer Verlag, New York.
Duffie J A and W A Beckman (2006 [1980]) Solar Engineering of Thermal
Processes, Wiley, New York.
Goswami Y, F Kreith and J Kreider (2000) Principles of Solar Engineering,
CRC Press, Boca Raton, FL.
Kalogirou S (2009) Solar Energy Engineering: Processes and Systems,
Academic Press, New York.
Larson R and West R (1996) Implementation of Solar Thermal Technology
(Solar Heat Technologies), MIT Press, Cambridge, MA.
Sayigh A (1978) Solar Energy Engineering, Academic Press, New York.
Winter C J, R L Sizmann and L L Vant-Hull (eds) (1991) Solar Power Plants,
Fundamentals, Technology, Systems, Economics, Springer Verlag, New
York.
2.13
References
Bejan A, G Tsatsaronis and M Moran (1996) Thermal design and optimization,
Wiley, New York.
Bergman T L, A S Lavine, F P Incropera and D P DeWitt (2011) Fundamentals of
Heat and Mass Transfer, 7th edn., Wiley, New York.
Blanco M, A Mutuberria, A Monreal and R Albert (2011) Results of the empirical
validation of Tonatiuh at Mini-Pegase CNRS-PROMES, Proceedings of SolarPACES 2011, Granada, Spain. See also http://code.google.com/p/tonatiuh/.
Buie D, A G Monger and C J Dey (2003) Sunshape distributions for terrestrial solar
simulations, Solar Energy, 74, 113–122.
Çengel Y and M Boles (2010) Thermodynamics: An engineering approach,
McGraw-Hill, New York.
Chaves J, (2008) Introduction to Nonimaging Optics, CRC Press, Boca Raton, FL.
Eck M, H Barroso, M Blanco, J-I Burgaleta, J Dersch, J-F Feldhoff, J GarciaBarberena, L Gonzalez, T Hirsch, C Ho, G Kolb, T Neises, J A Serrano, D Tenz,
M Wagner and G Zhu (2011) guiSmo: guidelines for CSP performance modelling
– present status of the SolarPACES Task-1 project, Proceedings of SolarPACES
2011, Granada, Spain.
Garcia P, A Ferriere and J Bezian (2008) Codes for solar flux calculation dedicated
to central receiver system applications: A comparative review, Solar Energy, 82(3),
189–197.
Gordon J (ed.) (2001) Solar Energy – The State of the Art, International Solar
Energy Society, James and James, London.
Ho C K (2008) Software and Codes for Analysis of Concentrating Solar Power
Technologies, Sandia report SAND2008-8053, Sandia National Laboratories,
Albuquerque.
Kopp G and J L Lean (2011) A new, lower value of total solar irradiance:
Evidence and climate significance, Geophysical Research Letters 38, L01706,
doi: 10.1029/2010GL045777.
Llorente García I, J L Álvarez and D Blanco (2011) Performance model for parabolic trough solar thermal power plants with thermal storage: Comparison to
operating plant data, Solar Energy, 85, 2443–2460.
© Woodhead Publishing Limited, 2012
Fundamental principles of concentrating solar power (CSP) systems 67
Lovegrove K, G Burgess and J Pye (2011) A new 500 m2 paraboloidal dish solar
concentrator, Solar Energy, 85, 620–626.
Meinel A B and M P Meinel (1976) Applied Solar Energy, An Introduction,
Addison-Wesley, Reading, MA.
Moran M J and H N Shapiro (2011) Fundamentals of Engineering Thermodynamics,
John Wiley and Sons, New York.
NREL (2012) System Advisor Model (SAM). https://sam.nrel.gov/, accessed 29 February 2012.
OpenFOAM Foundation (2011) OpenFOAM User Guide, http://www.openfoam.
org/docs/user/, version 2.1.0, accessed 29 February 2011.
Rabl A (1976) Comparison of solar concentrators, Solar Energy, 18, 93–111.
Rabl A and P Bendt (1982) Effect of circumsolar radiation on performance of
focusing collectors, Journal of Solar Energy Engineering, 104(3), 237.
Short W, D J Packey and T Holt (1995) A Manual for the Economic Evaluation of
Energy Efficiency and Renewable Energy Technologies, NREL/TP-462-5173,
National Renewable Energy Laboratory, Colorado.
Steinfeld A and M Schubnell (1993) Optimum aperture size and operating temperature of a solar cavity receiver, Solar Energy, 50, 1925.
Stoddard M C, S E Faas, C J Chiang and A J Dirks (1987) SOLERGY – A Computer
Code for Calculating the Annual Energy from Central Receiver Power Plants,
SAND86-8060, Sandia National Laboratories, California.
Walton G N (2002) Calculation of obstructed view factors by adaptive integration,
Technical Report NISTIR–6925, National Institute of Standards and Technology,
Gaithersburg, MD. See also http://view3d.sf.net/.
Winston R, W T Welford, J C Miñano and P Benítez (2005) Nonimaging Optics,
Elsevier Academic Press, Maryland Heights, MO.
© Woodhead Publishing Limited, 2012
3
Solar resources for concentrating
solar power (CSP) systems
R. M E Y E R, M. S C H L E C H T and
K. C H H AT BA R, Suntrace GmbH, Germany
Abstract: Direct sunlight or beam irradiance is the key resource for any
concentrating solar system. Beam irradiance has a significantly higher
variability in space and time in comparison to global irradiance, and
its measurement requires higher accuracy and attention. Therefore
uncertainty of beam irradiance is higher and solar resources must be
measured with great care. In order to get realistic long-term values,
satellite-derived values are taken into account in addition to ground
measurements to mitigate the high inter-annual variability. The shortterm variability of beam irradiance in terms of fluctuations should be
properly represented as CSP systems are sensitive to transient
conditions.
Key words: solar radiation, direct normal irradiance (DNI), irradiation,
pyrheliometer, rotating shadowband irradiometer (RSI), satellite-derived
irradiance, radiation measurement.
3.1
Introduction
A concentrating solar thermal power plant converts solar irradiance into
thermal energy, and ultimately into electrical power. This is done by means
of high temperatures and efficient heat-to-electricity conversion systems,
such as steam turbines. Conventional fossil-fuelled power plants actively
control the rate of heat generation, and thus can adapt energy production
to match demand. Usually they can be classified into peak-load, mid-load
and base-load categories, each having different flexibilities in start-up, load
change and shut down characteristics. These processes are manageable and
established in practice.
In solar thermal power systems, the controllable conventional fuel is
replaced with solar irradiance, a variable source of energy, which is not in
the control of the plant operator. Solar irradiance incident at ground level
depends mainly on the sun elevation angle, which defines its path length
through the atmosphere and on the constituents of the atmosphere. Clouds,
aerosols and water vapour play a dominant role that affects the amount of
solar radiation reaching the ground. The amount and fluctuation of these
atmospheric constituents obviously are beyond the direct control of the
68
© Woodhead Publishing Limited, 2012
Solar resources for concentrating solar power (CSP) systems
69
plant. To maximize the yields of a solar thermal plant, available solar irradiation must be used to the largest possible extent.
Acquiring information about typical solar irradiation available at potential plant sites is of high importance for planning of new solar plants. The
climatological average annual solar irradiance is most important in the
planning process, but the characteristic frequency distribution also plays an
important role. Both phenomena are summarized under the notion ‘solar
resources’, which is the subject of this chapter. The term ‘solar resources’
here does not include ‘solar irradiation forecasting’, which more precisely
is the process of predicting solar radiation conditions for certain time steps
in the future.
This chapter gives an overview of those solar radiation characteristics,
which are relevant for concentrating solar power (CSP) plants. It defines
and describes the most important terms and units. The use of satellites to
derive solar radiation data is discussed. Best practices to reach reliable
estimates of solar irradiance conditions for a specific CSP plant location are
explained.
As solar thermal power plants are very large installations and often
exposed to harsh desert environments, other meteorological parameters
also need to be analysed for specific project locations. In the worst case,
wind gusts might harm the large sail-like mirror structures. Ambient air
temperature and humidity affect cooling conditions and thus influence the
efficiency of the power cycle. These parameters are also required for sitespecific engineering and are summarized in Section 3.6 on auxiliary meteorological parameters. Section 3.7 gives practical recommendations for a
step-by-step approach along the process of project development.
3.2
Solar radiation characteristics and assessment of
solar resources
Solar resource assessments should describe characteristic solar radiation
conditions based on historic data for the assessed plant location, assuming
similar intensities and patterns for future time periods. Maps of solar
resources are of highest importance for site selection (see Chapter 4). Solar
resource characteristics in the form of time-series in high time-resolution
are an important requirement for site-specific engineering and optimization
of plant layout towards high yields with moderate technical effort. Wellproven site-specific solar resource assessment is usually an essential factor
during due diligence assessment of CSP projects. Actual solar forecasting
then is eventually needed on-line with CSP plant operation. Similar techniques to those used for assessing solar resource in the early project stages
are applied, but focus of this chapter is the assessment of solar resources.
© Woodhead Publishing Limited, 2012
70
Concentrating solar power technology
All CSP technologies need to concentrate sunlight. As optical concentration cannot be achieved based on diffuse light coming from various directions, only the direct beam irradiance component is relevant for CSP. Direct
beam irradiance is usually specified in reference to a tracking area oriented
normal towards the sun and hence is called direct normal irradiance (DNI).
In contrast global horizontal irradiance (GHI) includes both the direct
irradiance and the diffuse irradiance (related to the horizontal plane). As
non-concentrating photovoltaic (PV) can also utilize a substantial amount
of diffuse irradiance, GHI is closely related to the assessment of PV energy
yields, while DNI is applied for estimation of energy yields from CSP and
CPV (concentrating PV) plants.
The distribution of DNI across the globe is much less homogeneous than
GHI (see Plate I between pages 322 and 323). While GHI reaches relatively
good values in latitudes higher than 45° and is also quite high around the
equator in tropical climates, DNI is usually highest in the subtropics around
23°N or 23°S. These latitudes ±10° or around 1000 km from the equator are
the regions of the world where the highest DNI values are reached. Therefore these bands on both sides of the equator are also called the
‘sunbelts’.
Site-specific DNI data are difficult to obtain and often include a high
uncertainty. Therefore special care should be addressed to solar resource
assessment during project development. The long-term average of solar
irradiation and its variability need to be analysed for a reliable projection
of their availability in the future. This must carefully take into consideration
the uncertainties related to the derivation of solar irradiance data. When
evaluating the conversion of solar irradiance into thermal energy and then
into electricity, the fluctuation patterns of solar irradiation have transient
effects on plant performance and thus influence energy production. The
impact of these effects depends on system inertia, technical configuration
of the plant and operational and control strategies. To allow the detailed
assessment of these effects with sophisticated energy yield modelling tools,
the DNI data should be available in small time steps with high accuracy
and realistic frequency distribution.
The typical lifetime of a solar thermal power plant is expected to be 25
to 40 years. The financing of plants usually assumes a loan period within 15
to 25 years. Compared to global irradiance, the direct beam component
shows much more variability in space and time. DNI values obtained from
a data source vary from year to year and inter-annual variability can be
very extreme. This variability in annual averages of DNI is reflected in the
estimation of energy yield of CSP plants.
A general assumption in climatology is that when considering meteorological data spanning 30 years, weather conditions at the site are averaged
out and hence they can be used to calculate the long-term average of the
© Woodhead Publishing Limited, 2012
Solar resources for concentrating solar power (CSP) systems
71
meteorological conditions. Under the best possible condition, groundmeasured meteorological data covering the above-mentioned period should
be available. But in reality 30 years of measured data are available only for
few scientifically monitored locations, but not for potential locations of
commercial CSP plants. Lohmann et al. (2006) shows that for most sites in
the sunbelt, if data from at least 10 years are taken into account the
maximum deviation of DNI-averages falls below ±5% from the long-term
average.
Thus, for site qualification and yield evaluation of potential construction
sites for solar thermal power plants it is highly desirable to have reliable
historical data of direct irradiation, ideally for at least 10 years. However
for most potential plant locations only satellite data are available for such
period.
3.2.1
Important solar radiation terms
The term irradiance is used to describe the solar power (instantaneous
energy flux) falling on a unit area per unit time, i.e. in W/m2. The term irradiation is used to consider the amount of solar energy falling on a unit area
over a stated time interval such as a day or a year. In climatology, solar
irradiation typically is given in J/m2 summed over the period of a day,
whereas solar engineering solar irradiation is often reported in units of
kWh/m2.
According to the standard EN ISO 9488 (ISO, 1999), the symbol for the
irradiance is G and for the irradiation is H. However, often the same
symbols are used for irradiance and irradiation. Then they have to be differentiated by context or by the attached units.
When the term ‘irradiation’ is used, it is necessary to indicate the period
over which the irradiance is integrated, as it cannot be seen in the units.
Therefore it is better to avoid the term irradiation and use only irradiance.
Instead of the annual sum, the yearly average of irradiance can be indicated
in units of kWh/(m2a); instead of daily irradiation, the daily average irradiance may be indicated by kWh/(m2d). In the CSP industry, the most commonly used units for solar irradiation are kWh/(m2a) or kWh/(m2d) or W/
m2. The conversion of solar irradiance from one unit to another is shown
in Table 3.1. This assumes years of 365 days, neglecting the effect of leap
years. Following this, for example, a DNI of 2000 kWh/(m2a) ≈ 5.48 kWh/
(m2d) ≈ 228 W/m2 is equivalent to a daily mean of 19.7 MJ/m2.
In meteorology, average solar irradiance values usually consider full days
with 24 hours including nighttime. However, in engineering, sometimes only
the period when the sun is above the horizon is relevant for design purposes.
The latter has the advantage that average irradiance is closer to actual
intensity observed during the day. Irradiance averages only referring to
© Woodhead Publishing Limited, 2012
72
Concentrating solar power technology
Table 3.1 Conversion table for solar irradiance values related to 24
hour days
Unit
W/m2
kWh/(m2d)
kWh/(m2a)
MJ/(m2d)
Value
1
0.024
8.76
11.574
sunshine duration and not to intensity are more complicated to calculate
and less correlated to actual daily irradiation. Ambiguities can also occur if
sunrise and sunset times are not calculated astronomically assuming a flat
horizon, but from the actual period covering first and last visibility of the
sun, which depends on local topography. To avoid these problems and to
easily convert from average values to daily or annual values, it is recommended to refer to 24-hour average irradiance values.
The extraterrestrial solar irradiance is the solar radiation received at the
top of the Earth’s atmosphere. The extraterrestrial solar irradiance varies
slightly due to variations in energy emitted by the sun and variable distance
of the Earth from the sun in the course of the year. The solar constant I0 is
defined as the solar irradiance at the top of the earth’s atmosphere on a
plane normal to the direction of this radiation, when the Earth is at its mean
distance from the sun (149,597,871 km). Its measured average value in the
current period is 1366 W/m2 ± 0.6 W/m2 (ISO, 2007).
As illustrated in Fig. 3.1, the extraterrestrial irradiance is attenuated on
its way through the atmosphere. The unattenuated extraterrestrial irradiance, which is not absorbed or scattered by the atmosphere and reaching
the surface directly, is the direct irradiance Gb, which is related to the horizontal plane. According to ISO 9448 (1999) DNI is defined as the direct
irradiance on a plane normal to its angle of incidence. Direct beam irradiance strictly refers to non-scattered solar radiation. This comes from the
solar ‘disk’only, which covers a solid angle of around 0.5°. Due to measurement reasons ISO 9448 (1999) allows an acceptance angle of up to 6°
around the centre of the sun’s disk for measurement of DNI. On the other
hand, most radiative transfer algorithms define beam irradiance as only that
particular part of solar radiation which does not experience any scattering
– even forward-scattering within the 0.5°-cone of the sun is not considered
in this definition. Therefore, simulated direct irradiance is slightly lower
than the beam irradiance DNI. As most CSP technologies have a wider
acceptance angle than 0.5°, the DNI-definition related to the typically measured 5° cone is considered here.
The diffuse irradiance Gd (see Fig. 3.1) is the scattered irradiance that
reaches the ground. Scattering might occur by various processes in the
atmosphere, like Rayleigh scattering by air molecules or Mie scattering by
© Woodhead Publishing Limited, 2012
Solar resources for concentrating solar power (CSP) systems
e, I 0
anc
rre
rate
Ext
p
To
of
Absorption
≈1%
ere
ph
s
mo
Ozone
at
Air
Rayleigh scattering + absorption
≈15%
Air molecules
ma
ss
Zenith
ia
str
=2
Air mass = 1
adi
l irr
73
Aerosol
Scattering + absorption
variable, typically 15%
Scattering + absorption
very variable, 0–100%
Clouds
Absorption
≈15%
Water vapor
vel
a le
P
Se
3.1 The main processes influencing solar radiation in the atmosphere
and split into the major three components (global, direct, diffuse).
aerosol or cloud particles. Solar radiation reaching the surface, where it
partly gets reflected and backscattered, e.g. by clouds, also contributes to
diffuse irradiance. If albedo is high, e.g. in regions with snow or white sand
cover, the latter process can play a significant role.
Direct horizontal irradiance is related to DNI via the cosine of the solar
zenith angle. The sum of direct horizontal irradiance and the diffuse horizontal irradiance (DHI) results in the total irradiance or global horizontal
irradiance:
G = Gd + Gb = DHI + DNI/ cos(θ z )
[3.1]
where θz is the solar zenith angle (see Fig. 3.2).
3.2.2
Seasonal variation of global and beam irradiance
Knowledge of geometrical parameters of the solar trajectory is very important in solar energy applications. Every 365.25 days the Earth revolves
around the sun in an elliptical orbit with a mean Earth–sun distance of
149,597,871 km, defined as one astronomical unit. The plane of this orbit is
called the ecliptic plane. In this century, the maximum distance of Earth
from the sun of 152,101,100 km, which also is called aphelion, is reached by
© Woodhead Publishing Limited, 2012
74
Concentrating solar power technology
Z
L: longitude
f: latitude
ws: hour angle
d: declination angle
qz: zenith angle
qz
P
d
L
f
Equator
Y
ws
Sun
Prime meridian
X
3.2 Solar position in the terrestrial coordinate system for point P at a
location of approximately 15 degrees E 45 degrees N at approximately
13:00 UTC.
the Earth’s orbit on about 3 July. The perihelion, which is the minimum
Earth-sun distance, occurs on about 2 January, when the Earth is
147,101,100 km from the sun.
The Earth rotates about its own polar axis inclined to the ecliptic plane
by 23.45°, in approximately 24-hour cycles which produces day and night.
The tilt of this axis relative to the ecliptic plane produces the seasons as the
Earth revolves around the sun. To predict the direction of sun rays relative
to a point on the Earth, the solar time is used, which is dependent on local
longitude and is generally different from local clock time. Consequently, at
12:00 solar time, the sun is at its highest point in the sky and exactly due
south in the northern hemisphere, or north in the southern hemisphere. The
regional clock time is instead defined by politically defined time zones.
Accurate knowledge of the difference between solar time and the local
clock time is required for energy demand correlations, system performance
correlations, determination of true south, and tracking algorithms.
As seen in Fig. 3.2, the Earth’s rotation around its polar axis is described
by the solar hour angle ωs, which is the angular distance between the meridian of the observer, and the meridian, whose plane contains the sun. The
solar hour angle is zero at solar noon, since the meridian plane of the
observer contains the sun at this time, the sun is said to be ‘due south’ for
an observer in the northern hemisphere or ‘due north’ in the southern
hemisphere. The solar hour angle increases by 15° every hour and is calculated by:
Solar hour angle ω = 15° ⋅ Δt
© Woodhead Publishing Limited, 2012
[3.2]
Solar resources for concentrating solar power (CSP) systems
75
Any location on the surface of the Earth can be defined by the combination
of a longitude angle L and a latitude angle φ. The angular distance between
the projections of meridian of the local point and the prime meridian on
the equatorial plane is defined as the longitude.The angle between the line
from a point on the Earth’s surface to the centre of the Earth and the
Earth’s equatorial plane is the latitude. The equator at 0° latitude designates
the intersection of the equatorial plane with the surface of the Earth. The
Earth’s surface at 90° latitude (north pole) and −90° latitude (south pole)
is intersected by the earth’s axis of rotation.
Figure 3.2 also shows the declination δ, the angle between the Earth’s
equatorial plane, which is the plane that includes the Earth’s equator, and
the line of the centre of the Earth and the sun. On 21/22 June at noontime
the sun is at its highest point in the sky in the northern hemisphere, and its
lowest in the southern hemisphere, with a declination of +23.45°, because
at this time of the year the Earth’s equatorial plane is inclined 23.45° to the
Earth–sun line. This condition marks the beginning of summer in the northern hemisphere and is called summer solstice.
About three months later, i.e. on 22/23 September, a line from the Earth
to the sun lies on the equatorial plane and the declination is zero; this condition is called an equinox. At this time the sun at the equator is directly
overhead at noontime. Anywhere on the Earth during an equinox, the time
during which the sun is visible (daytime) is exactly 12 hours. There are two
such conditions during a year. There are the autumnal equinox around 22/23
September, marking the start of autumn and the vernal equinox around
20/21 March, marking the beginning of spring.
On about 21/22 December the winter solstice occurs and marks the point
where the equatorial plane is tilted relative to the Earth-sun line such that
the northern hemisphere is facing away from the sun. Consequently, at
noontime the sun is at its lowest point in the sky, meaning that the declination is at its most negative value. Northern winter declination angles are
negative by convention.
For CSP systems it is essential to track the sun exactly during the day
and also from one day to another. For this purpose the sun is observed from
a position on the Earth’s surface. It is of interest to define the sun’s position
relative to a local coordinate system (see Fig. 3.3). A zenith line (straight
up) and a horizontal plane containing a north–south line and an east–west
line are the conventional Earth-surface based coordinates.
The actual position of the sun from a point can be defined by two angles:
solar azimuth γs and solar elevation or solar height hs. The solar height (or
solar elevation) is defined as the angle between the centre of the sun and
a horizontal plane containing the observer. Solar height alternatively can
be indicated in terms of the solar zenith angle θz, which is simply the
complement of the solar elevation angle.
© Woodhead Publishing Limited, 2012
76
Concentrating solar power technology
Sun
Zenith
N
qz
h
W
E
P
s
qz: zenith angle
h: elevation angle
s: azimuth angle
S
3.3 Solar position viewed from a point P on the Earth’s surface in a
local coordinate system.
θ Z = 90° − hs
[3.3]
The disadvantage of the term sun height is that it requires a clear definition
of the unobstructed flat horizon. It is more difficult to fix this plane compared to the direction of the zenith, which is defined by the vertical line,
which aligns at most places better than 0.1° with the plumb line.
The solar azimuth is the second angle used to define the exact position
of the sun. Solar azimuth angle is defined as the angle between the projection of sun’s centre onto the horizontal plane and due south direction.
According to ISO 9488 (1999) it is defined as 0° at solar noon and increases
thereafter, when the sun position goes toward west. Before noon towards
east it is negative reaching 0° at solar noon as the day progresses. Alternatively, it is often indicated clockwise from geographic north to the projection of the sun’s centre onto the horizontal plane.
For solar energy system design and operation, it is important to be able
to calculate both solar angles at any time for any location on the Earth. For
this task many different algorithms have been proposed (e.g. Spencer
(1971); Michalsky (1988); Blanco et al., (2001); Reda and Andreas (2004)).
Simple algorithms only consider the geometrical position of the sun. Others
also take into account the angular deviation due to refraction in the atmosphere. According to the review paper of Lee et al. (2009) it is recommended to use, for easy and fast applications, Grena (2008) or for
applications requiring higher accuracy, the computational more expensive
algorithm of Reda and Andreas (2004).
© Woodhead Publishing Limited, 2012
Solar resources for concentrating solar power (CSP) systems
3.2.3
77
Influence of atmospheric constituents on direct
beam irradiance
Gases, liquid and solid particles and clouds attenuate the intensity of extraterrestrial solar radiation traversing through the Earth’s atmosphere. Three
groups of atmospheric constituents determine the interaction of solar radiation with the Earth’s atmosphere:
gases (air molecules like ozone O3, carbon dioxide CO2, oxygen O2 and
water vapour H2O),
• solid and liquid particles (aerosols),
• clouds (condensed water in the form of droplets or ice particles).
•
The available solar radiation at the Earth’ surface is primarily linked to the
length of the optical path through the atmosphere. This is based on solar
position above the horizon as described above. The topography of the location defines the elevation of a point, as well as shadowing by neighbouring
terrain features. These geometrical factors can be modelled to a high level
of accuracy. The attenuation by gas constituents is mainly caused by
absorption and by Rayleigh scattering. The attenuation by aerosols is dominated by Mie scattering and also some absorption. A summary of all noncloud effects can be described by the Linke turbidity TL. It indicates the
optical density of a hazy cloud-free atmosphere in relation to a clean and
dry atmosphere. TL is the number of clean dry air masses that would result
in the same attenuation of radiation as the actual hazy and humid air.
Because of the dynamic nature of turbidity, its calculation and subsequent
averaging leads to some generalization. There are seasonal changes of turbidity, in which the lowest values in many regions appear in winter and
springtime and higher values in summer. The values of turbidity differ from
place to place in a similar degree of magnitude. Differences are related to
the terrain elevation, the intensity of industrialization and urbanization.
However, the strongest attenuation of solar radiation at the ground level
is from clouds. For determining the influence of clouds on solar radiation,
detailed information regarding geometrical thickness, position and number
of layers of clouds, as well as their optical properties are required. Therefore, in solar energy the attenuation by clouds is often estimated by simple
empirical techniques.
3.2.4
Spectral characteristics of solar radiation
Solar radiation is electromagnetic radiation emitted by the sun, in which
approximately 99% of the solar radiation incident on the Earth’s surface is
encompassed within the wavelength range from 0.3 μm to 3.0 μm. This
wavelength range is called the solar range. Figure 3.4 shows its spectral distribution outside the atmosphere and at sea-level after molecular
© Woodhead Publishing Limited, 2012
78
Concentrating solar power technology
Spectral irradiance (W/(m2 nm))
2.0
Extraterrestrial spectrum
DNI at sea level (AM = 1.5)
1.8
1.6
1.4
1.2
O2
1.0
0.8
0.6
H2O
0.4
CO2
0.2 O3
0.0
0
500
1000
1500 2000 2500
Wavelength, λ (nm)
3000
3500
3.4 Spectral distribution of sunlight and molecular absorption at sea
level with standard atmosphere.
absorption. The sun’s spectrum is dominated by the emission of the sun’s
outer spheres. It can be approximated by a black body with a surface temperature of approximately 5,778°K or around 6,000°C. Ozone is the main
absorption factor in the ultraviolet spectral range (100 nm to 400 nm, according to ISO 21348 (2007)) and, with less impact at 612 nm. Water vapour is
another absorption factor with impact at wavelengths above 500 nm, but
with increased impact above 1000 nm. At this wavelength, carbon dioxide is
also a strong absorption factor, at a similar wavelength to water vapour.
3.3
Measuring solar irradiance
Measuring the sun’s energy incident on the Earth’s surface is one of the
most difficult field measurement exercises. The measurement technology
applied today is based on an energy conversion process whereby electromagnetic radiant energy is converted into another form of energy, which
can be detected by measurements. Conversion into an electric signal is
preferred.
Solar instruments with a hemispherical (180°) field of view are called
pyranometers. In contrast are pyrheliometers, instruments using only a
narrow field of view (typically 5°). These are designed to measure the radiation coming from the solar disc and the adjacent region around the sun
(circumsolar). Consequently, a pyrheliometer must be accurately tracking
the sun to keep it properly oriented.
© Woodhead Publishing Limited, 2012
Solar resources for concentrating solar power (CSP) systems
(a)
(b)
79
(c)
3.5 Thermopile pyrheliometer instruments (a) Kipp & Zonen CHP1 (b)
Eppley solar tracker with two arms for pyrheliometers and (c) Eppley
NIP.
The sensor principles employed today either follow the thermoelectric
or the photoelectric effect. Both have specific advantages and shortcomings.
3.3.1
Thermal sensors
With thermal solar radiation sensors, the radiant energy is initially converted into thermal energy by means of a black absorbing surface, and then
into an electrical signal by a thermopile. This electrical output can be measured with a voltmeter.
As an example, various thermal pyrheliometers which are designed to
measure DNI directly are shown in Fig. 3.5. Alternatively, if measured by
pyranometers, DNI can be derived indirectly by simultaneous measurement
of global horizontal irradiance and diffuse irradiance, with arithmetic
deduction of values via equation [3.1]. Figure 3.6 shows examples of thermopile pyranometers designed to measure hemispherical irradiance.
Ideally, the black surface of thermopile instruments should absorb like a
perfect black body, which is a body fully absorbing and also emitting radiation of all wavelengths. Such a photon trap converts all radiation into heat,
allowing the conversion of solar radiation intensity into a temperature
signal, which then can be measured for example through a thermopile
element.
Such a thermopile element consists of a large number of thermocouple
junction pairs electrically connected in series. A thermocouple consists
of two dissimilar metals connected together. The absorption of thermal
© Woodhead Publishing Limited, 2012
80
Concentrating solar power technology
(a)
(b)
3.6 Examples of thermopile pyranometers (a) Kipp & Zonen CHP21
and (b) Eppley PSP.
radiation by the active (or ‘hot’) thermocouple junctions increases its temperature to T1 and the reference (or ‘cold’) junction is kept at a fixed temperature T2. The differential temperature between the active junction and
a reference (‘cold’) junction produces an electromotive force directly proportional to the differential temperature created. This effect is called the
thermoelectric effect. The magnitude and direction of the electromotive
force is affected by the type of metal and the temperature difference
between the hot and cold ends. The relationship between the temperature
difference and the output voltage of a thermocouple is nonlinear and is
approximated by a polynomial interpolation.
Due to the functional principle of thermopile radiometers, they are sensitive in a wide spectral range. Around 97–98% of the total irradiance energy
is absorbed by the thermal detector (Fig. 3.7). A disadvantage of thermopile
pyranometers in comparison to photoelectric pyranometers is a higher
price and frequent soiling of the glass dome (Pape et al., 2009).
3.3.2
Photoelectric sensors
Photoelectric instruments convert the radiant energy directly into
electrical energy by a photodiode. A photodiode is usually a silicon semiconductor with p-i-n structure or p-n junction. Photodiodes can be used
under either reverse bias (photoconductive mode) or zero bias (photovoltaic mode).
In zero bias, solar radiation incident on the diode causes a current across
the device, leading to forward bias, which in turn induces ‘dark current’ in
the opposite direction to the photocurrent. This is called the photovoltaic
© Woodhead Publishing Limited, 2012
81
Thermopile sensor
1.4
Photoelectric Si-sensor
DNI at sea level (AM = 1.5) 1.2
100
90
80
70
60
50
40
30
20
10
0
1.0
0.8
0.6
0.4
0.2
Spectral irradiance (W/(m2 nm))
Relative response (%)
Solar resources for concentrating solar power (CSP) systems
0.0
0
500
1000
1500
2000
2500
Wavelength, λ (nm)
3000
3500
3.7 Solar irradiance spectrum at the Earth’s surface and typical
pyranometer response functions (dashed line = thermopile instrument;
dotted line = silicon photodiode).
(a)
(b)
3.8 Photoelectric pyranometer LI-COR LI-200SZ (a) front view and (b)
top view.
effect, which is the basis for solar cells. Figure 3.8 shows the photoelectric
pyranometer LI-200SZ by LI-COR (2005).
The response of pyranometers, which measure the irradiance with a photodiode, have a much narrower spectral sensitivity compared to that measured by thermopiles (Fig. 3.7). Its spectral response typically is in the range
between 0.4 and 1.2 μm and it is not uniform. Therefore, narrow-tobroadband corrections need to be applied to derive the full solar range.
Ideally, these consider the spectral effect of lower air masses and also that
of various contents of atmospheric trace gases and aerosols. DNI can be
© Woodhead Publishing Limited, 2012
82
Concentrating solar power technology
(a)
(b)
3.9 RSI instruments (a) RSR2 of Irradiance Inc and (b) RSP4G of
Reichert GmbH.
determined from a single photoelectric pyranometer, if it is assembled in
an arrangement that periodically blocks the direct beam radiation and
causes it to measure global irradiance and diffuse irradiance alternately.
Rotating shadowband irradiometers (RSIs) usually employ a photoelectric radiometer (Fig. 3.8) to measure incident solar radiation. As shown in
Fig. 3.9 the pyranometer is mounted on a ‘head unit’ apparatus, which
permits unobstructed measurement of global horizontal irradiance (GHI)
and to measure horizontal diffuse irradiance (DHI) by means of a motordriven shadowband which periodically blocks the direct beam component.
RSI operation is realized by a program code, which drives a motor control
module. The code is usually run by a datalogger, which processes and collects the measured values. Typically once every 30 seconds the shadowband
rotates over the photodiode, taking approximately one second for this
motion. During this period the photoelectric pyranometer signal is sampled
about 1,000 times and when the sensor is completely shaded from the sun
by the shadowband the lowest pyranometer readings occur. During this
short moment, the photoelectric pyranometer measures only the diffuse
irradiance. The software detects the DHI from all values, finding the average
of minimum values. Finally, DNI is calculated from measurements of
GHI and DHI, and the sun’s zenith angle using relation mentioned in
equation [3.1].
DNI measurements derived by RSIs are not reaching the exact same
accuracy compared to pyrheliometer measurements, as long as the pyrheliometer tracking is accurate and the measurement device is maintained
and properly cleaned. Soiling of measurement devices is often an issue with
pyrheliometers, as these are easily affected by dust and dirt accumulation
© Woodhead Publishing Limited, 2012
Solar resources for concentrating solar power (CSP) systems
83
and need daily and careful cleaning. In practice, often the cleaning of pyrheliometers is not practical as this is difficult to carry out in remote desert
locations – the typically preferred areas for CSP application. RSI type
instruments on the opposite are less susceptible to soiling and thus require
very little attendance and maintenance. They are also less expensive compared to a pyrheliometer station with a solar tracker. For maximum data
reliability, solar monitoring stations with both sorts of instruments are typically configured with redundant sensors, which also allows comparison of
both sensors regarding parallel measurement of the same DNI intensity.
3.4
Deriving solar resources from satellite data
In cases where ground-based measurements are not available, satellitederived solar radiation values are used. Satellites measure reflected radiation from the Earth’s surface in several wavelength bands. Known albedo
values by location and complex models and algorithms can be used to
determine global, diffuse and direct beam irradiance components. Raw data
are available from various satellite operators and these in turn are processed by several different organizations providing solar resource satellite
data services. In some cases these are commercial services and in other
cases, research or government. One of the best known data set is provided
by NASA (http://eosweb.larc.nasa.gov/sse/).
The NASA website offers DNI data to the public without charge for any
location across the globe. The data are in the form of monthly averages and
were derived from 22 consecutive years of satellite data with a basic resolution of approximately 30 km, which has been processed to a 1° × 1° geographical grid.
All satellite data providers have different temporal and spatial coverage,
different temporal and spatial resolution, use different algorithms and different inputs. As a result, solar radiation values derived for a given place
from different satellite providers can differ significantly.
There are many advantages associated with satellite-derived values:
•
satellites have high spatial resolution covering most of the sites of
interest
• the temporal coverage of meteorological satellites is quite long; some
data date back to the late 1970s, and more or less continuous coverage
in some regions can be reached since the 1980s. Since satellites cover
such long terms of data, the historic data basis can be applied to estimate
future irradiation, and thus can serve as basis for planning and sizing of
solar systems
• satellite-derived solar radiation values (real-time data) can be utilized
for monitoring and managing dispersed solar power in the grid.
© Woodhead Publishing Limited, 2012
84
Concentrating solar power technology
Meteo input
(DNI; T; RH; wind)
Solar plant
component data
Operation
strategies
Yield simulation
Financial yields
Energy yield results
3.10 Flow chart showing the energy yield evaluation process.
The derivation of solar irradiation values from satellite data is realized by
radiative transfer calculations. Irradiation is calculated from infrared and
visible channels of satellites with a nominal spatial resolution of 3 km ×
3 km at ground level and 5 km × 5 km at ground level for the second and
first generation of Meteosat, respectively. The annual sums of satellitederived data can coincide with ground-measured data within a range of
+/− 5%. However, in some cases deviations of around 20% have been
observed (Gueymard, 2010). For shorter time periods like single months,
days or hours the deviation of satellite-derived and measured solar irradiance values can increase significantly (Zelenka et al. 1999).
Due to insufficient accuracy of satellite data, precise ground-measured
data are essential for validation and calibration of long-term satellitederived data. The minimum measurement period should cover a duration
of one year, so that at least one complete seasonal cycle can be
evaluated.
Figure 3.10 shows the simplified structure of the power plant energy yield
evaluation process, with obtaining and processing of the solar irradiance
measurement being integrated.
3.5
Annual cycle of direct normal irradiance (DNI)
Whilst annual average DNI levels are a good indicator of potential annual
generation levels, there is often large variation of solar radiation values
observed in different seasons. It is important that this be taken into account
during project feasibility studies, as the varying level of average output
with seasons through the year can affect plant economics and off-take
agreements. Plate II (between pages 322 and 323) illustrates the global
distribution of DNI at different times of the year. Obviously the strongest
effect is the underlying variation in DNI when a hemisphere is further from
© Woodhead Publishing Limited, 2012
Solar resources for concentrating solar power (CSP) systems
85
the sun during winter or closer to the sun during summer. This variation
increases with growing distance to the equator. However, in addition to this,
there are (seasonal) variations in cloud cover, which are very locationspecific. For example, in tropical regions near the equator, monsoon-type
weather cycles dominate the pattern.
3.6
Auxiliary meteorological parameters
Solar radiation and particularly DNI is the major meteorological parameter
that influences energy yields of CSP plants. The influence of other meteorological parameters like ambient temperature, wind speed, relative humidity, etc., is minor as long as the variations are not too extreme. Therefore,
such parameters are termed auxiliary. The auxiliary meteorological parameters that affect the performance of CSP plants include ambient temperature and relative humidity. These parameters, for example, have influence
on the operating conditions and, thus, the efficiency of the steam turbine
cooling system. Ambient temperatures also have an effect on thermal losses
from receivers.
If wind speed regularly exceeds design limits, it may lead to higher frequency of solar field safety shutdowns (moving to the stow position), reduce
performance due to increased thermal and optical losses (shattering of
mirrors), potential damage of system components due to higher forces on
structures and motors, reducing plant availability and increasing operational expenditures (OPEX) (Chhatbar and Meyer, 2011).
3.6.1
Air temperature
The ambient temperature has two contrasting effects: one on the
efficiency of the solar field and other on the efficiency of power block.
The efficiency of the solar field depends on the convective losses of the heat
transfer fluid (HTF) and the collectors to the ambient air. These losses are
dependent on ambient temperature. The lower the ambient temperature,
the higher the losses and vice versa. In addition, the efficiency of the power
block is indirectly a function of the ambient temperature. The overall efficiency of power block is dependent on the condenser efficiency. For wet
cooling, the efficiency of the condensers increases with decreasing wet bulb
temperature, which is a function of ambient temperature and relative
humidity, and vice versa.
3.6.2
Humidity
The efficiency of the wet cooling system decreases with increasing relative
humidity, which in turn results in reduced efficiency of the power block. As
© Woodhead Publishing Limited, 2012
86
Concentrating solar power technology
a result the overall energy yield of CSP plants is affected by changing relative humidity conditions.
3.6.3
Wind speed
At higher wind speeds, optical losses in the solar field increase because of
distortion of the geometry of collectors and this reduces efficiency. Moreover, convective heat losses in the solar field also increase with increasing
wind speed, which further reduce the efficiency. Too high wind speed conditions might also lead to plant shutdown in order to protect the mirrors
from being damaged, further reducing performance. Thus, in general it is
assumed that the energy yield decreases with increasing wind speed. As a
result, it is necessary to take these parameters into consideration while
assessing the solar resources at the site of interest for a CSP project. Moreover, once a meteorological station is set up to measure solar radiation,
additional equipment should also be used to measure the corresponding
parameters.
3.7
Recommendations for solar resource assessment
for concentrating solar power (CSP) plants
At present there is no standard procedure for processing solar radiation
data or a set of procedures to be followed for solar resource assessments.
As a result, and due to pressing deadlines for projects, project developers
do not always follow a strict approach for solar resource assessments. The
majority of the decisions for site selection of CSP plants are still based on
the usually rough initial site assessments. In such cases, the solar radiation
conditions such as annual cycle of DNI, frequency distribution of DNI,
inter-annual variability and the uncertainty have not been assessed in detail,
and as a result the current knowledge of solar resources available at the
site stays limited.
Based on practical experience and analysis of various data products and
methods, best practices to achieve high quality assessments with reasonable
effort are proposed. Following Meyer (2010) the recommended procedure
is as follows:
•
•
In the first place, when no detailed assessment of a project location has
been made, multiple satellite-based sources based on average values,
and if available, ground measurements, should be taken into consideration. At this point, long-term average values of DNI from these sources
may suffice.
Calculate a quality-weighted best average (Meyer et al., 2008) and determine the resulting uncertainty by Gaussian error propagation.
© Woodhead Publishing Limited, 2012
Solar resources for concentrating solar power (CSP) systems
Country/
regional
analysis
Prefeasibility
study
GIS-based
solar maps
Multiple satellite
sources
Feasibility
study
Site
qualification
Long-term
time-series
Measurements
87
Due
diligence
Expert opinions
3.11 Recommended steps for solar resource assessment in CSP
plants.
•
Once a project site has been confirmed for further development of the
CSP project, a suitable measurement station should be installed at the
site in order to reach the 1 year of measurement data in parallel with
the project approval process.
• Determine the long-term best estimate for the specific project site (in
1 km × 1 km resolution) and based on as many years of satellite data as
available. For DNI, a minimum of 10 consecutive years should be considered, because inter-annual variability is high.
• Multiple site-specific data sets from satellites and ground-based measurements, with overlapping time periods can provide independent
information and increase accuracy when combined. Only reliable data
sets shall be considered. The uncertainty of the data is determined individually for each data set.
Site specific satellite data time series are required to reflect the long-term
history of irradiation and must be adapted based on overlapping time
periods between different satellite and measurement data sets. To be sitespecific, at least a 10 km × 10 km spatial resolution is required, preferably
it should be 1 km × 1 km resolution.
The accuracy and reliability of the satellite data provider is important as
random and systematic errors can result in data of poor quality. Satellite
providers often have a regional focus, so that one provider is not always
recommended for all global regions. The current state of satellite-derived
methods leaves room for improvement. Also ground-based measurements
can be of poor quality depending on the individual maintenance of the
measurement station over the complete measurement period. Therefore,
when a location is seriously considered during the site qualification phase
the satellite-derived time series should be overlapped with, and adapted to
the ground-measured data, e.g., using the procedure of Schumann et al.
(2011). From the long-term satellite data, which are corrected with the
measurement data, the average DNI can be derived, representing the P50
value. P50 means, that the value should be exceeded in 50% of all years.
From the corrected satellite data the climatological average of DNI can also
be derived.
© Woodhead Publishing Limited, 2012
88
Concentrating solar power technology
For the creation of most reliable CSP-specific typical meteorological
years (TMY), site-specific measurements are of great value. According to
Hoyer et al. (2009) data selection for TMY-building should be done in a
way that for each month data are selected, which are closer than 5% to the
climatological average. The annual average in the TMY should be within
1% of the climatological average.
Finally, for bankable expert opinions on meteorological conditions for
CSP, usually data with a lower probability of underperformance are required
to satisfy conservative approaches from banks and lenders. So that in addition to the P50 TMY additional TMYs based on a more conservative
approach are required to assess the risk of lower irradiance values. For this
purpose typically either, e.g., P70 or P90 data sets are derived. These represent the data of a typical year, which values would be exceeded in 70%
or 90% of all years respectively.
Alternatively, a risk assessment using performance simulation results
based on several good and bad years could give sufficient comfort to the
banks. These time series can be derived from satellites but should also be
adapted to site-specific characteristics based on ground-measured data. This
approach would allow a more detailed assessment of the influence of variable DNI conditions.
Also from processing of at least 10 years of data, P70 or P90 values could
be derived, which usually are used as basis to calculate the financial base
case for a project. Compared to the simpler approach of only using P70 or
P90 years, the advantage of using multiple years is that on one side the
effects of meteorological variability and uncertainty, and on the other hand
the effect of uncertainties resulting from technical parameters (describing
the plant or the uncertainty of the performance simulation models) can be
assessed in more detail.
3.8
Summary and future trends
Solar resource data are currently available from different data providers
whose values differ from one another for the same geographic location. At
present there is no standard procedure for processing solar radiation data
or a set of procedures to be followed for solar resource assessments. Good
inter-comparable benchmarking of satellite-derived DNI products for sites
in CSP regions is an important missing link. Further improvement of measurements seems feasible and a clearer definition of processes will lead the
way to standardization of the overall task. Currently, often pyrheliometer
stations are applied, but are often not properly maintained and thus data
quality is reduced against the ISO standardized accuracy. On the other
hand, rotating shadowband pyranometers are also applied to derive DNI
for CSP project qualification and deliver more reliable data from remote
© Woodhead Publishing Limited, 2012
Solar resources for concentrating solar power (CSP) systems
89
and unattended locations, but are not yet ISO standard. Standardizing of
calibration and application of such instruments would be of benefit to the
industry. However, efforts are being made towards standardization and
integration of procedures for data bankability under the Task 46 of Solar
Heating and Cooling programme of the International Energy Agency. This
subtask mainly focuses on improving the procedures for measurement of
solar radiation for improving the accuracy. Optimization and standardizing
of procedures for combining satellite-derived long-term data sets with
ground-measured data is foreseen for sound planning and risk analysis of
large-scale solar energy projects. Moreover, efforts are also being made
towards benchmarking of satellite-derived data (Meyer et al., 2011).
The realm of resource forecasting is becoming more important for plant
dispatch as higher penetration of solar power is reaching the electric grid
systems. An accurate forecast could increase grid stability and solar plant
operator profits by optimizing energy dispatch into the time periods of greatest value. The accuracy of the available information can be improved by:
• additional meteorological measurement stations optimized for DNI in
the areas of interest for CSP and their proper operation and
maintenance
• improvement of the temporal resolution
• better availability of atmospheric data and higher accuracy for input
into satellite models
• improvement of satellite algorithms
• standardization of the procedures for measurements with RSI type
instruments and their calibration, input files and methodology for certification of solar resource assessment.
All measures described in this chapter mainly aim to derive as good as
possible the solar resources at CSP sites. An important prerequisite for all
introduced methods is that climate change will not significantly affect the
availability of DNI. In the course of the 21st century, however, it is expected
that there will be strong changes of regional climate, which then would also
affect the average DNI conditions. For the long-term operation of CSP this
might have some influence over its typical lifetime of more than 25 years.
Even though Lohmann et al. (2006) shows that during the 21 years from
1984 to 2004 no significant changes of DNI have occurred, it is likely that
the rate of climate change is increasing. Thus, additional research towards
this is recommended.
3.9
References
Benedikt, P., Batlles, J., Geuder, N., Piñero, R.Z., Adan, F., Pulvermueller, B. (2009)
Soiling impact and correction formulas in solar measurements for CSP projects.
In Proceedings of SolarPACES 2009 Symposium, Berlin, Germany.
© Woodhead Publishing Limited, 2012
90
Concentrating solar power technology
Blanco-Muriel, M., Alarcón-Padilla, D.C., López-Moratalla, T., Lara-Coira,
M. (2001) Computing the solar vector. Solar Energy 70, 431–441.
Chhatbar, K., Meyer, R. (2011) The influence of meteorological parameters on the
energy yield of solar thermal plants. In Proceedings of the SolarPACES Symposium, Granada, Spain, 20–24 September.
Grena, R. (2008) An algorithm for the computation of the solar position. Solar
Energy 82, 462–470.
Gueymard, C. (2010) Variability in direct irradiance around the Sahara: are the
modeled datasets of bankable quality? SolarPACES Conference, Perpignan,
France.
Hoyer, C., Hustig, F., Schwandt, M., Meyer, R. (2009) Characteristic meteorological
years from ground and satellite data. In Proceedings of SolarPACES 2009 Symposium, Berlin, Germany.
ISO (1999) EN ISO 9488, Solar energy – Vocabulary, International Organization
for Standardization.
ISO (2007) EN ISO 21348-2007, Space environment (natural and artificial) –
Process for determining solar irradiances, International Organization for
Standardization.
Lee, C.Y., Chou, P.-C., Chiang, C.-M., Lin, C.-F. (2009) Sun tracking systems: a review.
Sensors 9, 3875–3890; doi:10.3390/s90503875.
Lohmann, S., Mayer, B, Meyer, R. (2006) Long-term variability of global and direct
solar irradiance for solar energy applications. Solar Energy 80, 1390–1401.
Meyer, R. (2010) Recommendations for bankable meteorological site assessments
for solar thermal power plants. In Proceedings of the SolarPACES Symposium,
Perpignan, France, September.
Meyer, R., Butron, J.T., Marquardt, G., Schwandt, M., Geuder, N. Hoyer-Klick, C.,
Lorenz, E., Hammer, A. Beyer, H.G. (2008) Combining solar irradiance measurements and various satellite-derived products to a site-specific best estimate. In
Proceedings of the SolarPACES Symposium, Las Vegas, USA, 4–7 March.
Meyer, R., Gueymard, C., Ineichen, P. (2011) Standardizing and benchmarking of
model-derived DNI data products. In Proceedings of the SolarPACES Symposium, Granada, Spain, 20–24 September.
Michalsky, J. (1988) The astronomical almanac’s algorithm for approximate solar
position (1950–2050). Solar Energy 40, 227–235.
Pape, B., Batlles, J., Geuder, N., Zurita Piñero, R., Adan, F., Pulvermueller, B. (2009)
Soiling impact and correction formulas in solar measurements for CSP project.
Proceedings of SolarPACES 2009 Symposium, Berlin, Germany.
Reda I., Andreas, A. (2004) Solar position algorithm for solar radiation applications.
Solar Energy 76, 577–589.
Schumann, K., Beyer, H. G., Chhatbar, K., Meyer, R. (2011) Improving satellitederived solar resource analysis with parallel ground-based measurements. In Proceedings of the ISES Solar World Congress, 29 August–1 September, Kassel,
Germany.
Spencer, J.W. (1971) Fourier series representation of the position of the sun. Search
2, 172–173.
WMO (2006) Secretariat of the World Meteorological Organization, CIMO Guide
to Meteorological Instruments and Methods of Observation, Preliminary 7th edn,
WMO-No.8, Geneva, Switzerland.
Zelenka, A., Perez, R., Seals, R., Renné, D. (1999) Effective accuracy of satellitederived hourly irradiances. Theoretical and Applied Climatology 62, 199–207.
© Woodhead Publishing Limited, 2012
180°
120° W
60° W
0°
60° E
120° E
180°
60° N
60° N
30° N
30° N
0°
0°
30° S
30° S
60° S
60° S
180°
120° W
60° W
0°
60° E
120° E
60° N
180°
60° N
30° N
30° N
0°
0°
30° S
30° S
60° S
60° S
>3000 3000 2800 2600 2400 2200 2000 1800 1600 1400 1200 1000 800 <800
>8.22 8.22 7.67 7.12 6.58 6.03 5.48 4.93 4.38 3.84 3.29 2.74 2.19 <2.19
>342 342 319 297 274 251 228 205 182 160 137 114
91 <91
kWh/m2/a
kWh/m2/d
W/m2
Plate I World map of long-term global horizontal (GHI, top) and direct
normal (DNI, bottom) irradiance derived by Suntrace from DLR-ISIS
(Lohmann, 2006), NASA-SSE version 6.1 and adaption to elevation
and local measurements.
© Woodhead Publishing Limited, 2012
© Woodhead Publishing Limited, 2012
30° S
0°
30° N
30° S
0°
30° N
180°
180°
60° W
60° W
0°
0°
60° E
60° E
120° E
120° E
180°
180°
60° S
60° N
60° S
60° N
30° S 30° S
0° 0°
30° N 30° N
30° S 30° S
0° 0°
30° N 30° N
60° S
60° N
60° S
60° N
180°
180°
120° W
120° W
>3000 3000 2800 2600 2400 2200 2000 1800 1600 1400 1200 1000
>8.22 8.22 7.67 7.12 6.58 6.03 5.48 4.93 4.38 3.84 3.29 2.74
>342 342 319 297 274 251 228 205 182 160 137
114
120° W
120° W
0°
0°
800 <800
2.19 <2.19
91
<91
60° W
60° W
120° E
120° E
kWh/m2/a
kWh/m2/d
W/m2
60° E
60° E
180°
180°
60° S
60° N
60° S
60° N
30° S
0°
30° N
30° S
0°
30° N
Plate II World map of long-term seasonal averages of DNI: Northern hemisphere winter months (December to February in
upper left), spring (upper right), summer (lower left) and autumn (lower right). Maps derived by Suntrace from DLR-ISIS
(Lohmann, 2006), NASA-SSE version 6.1 and adaption to elevation and local measurements.
60° S
60° N
60° S
60° N
4
Site selection and feasibility analysis for
concentrating solar power (CSP) systems
M. S C H L E C H T and R. M E Y E R, Suntrace GmbH, Germany
Abstract: This chapter aims to provide an overview of the processes of
site selection and feasibility analysis for concentrating solar power (CSP)
projects and the challenges involved. It describes the aspects considered
in an iterative pre-feasibility analysis and a fully-fledged feasibility study.
These include: solar irradiance, site characteristics and infrastructure
connections, cost of installation and market and political environment.
All information gathered is considered as part of the overall project
development process. Specific conclusions are drawn for each project
depending on its boundary conditions and strategic goals.
Key words: concentrating solar power (CSP), site assessment, site
selection criteria, feasibility analysis, project development, project
viability, risk mitigation.
4.1
Introduction
Site selection and feasibility analysis are in principal two successional, independent tasks. The site selection process for concentrating solar power
(CSP) technology should lead to the identification of a potential site, then
a decision needs to be made as to the most suitable technical concept for
the project. When choosing the technical concept, one has to consider the
project economics, which depend on the cost of the technology, the financing conditions and, in particular, the revenue generated from energy sales.
National or regional government guidelines and rules may restrict size of
plants and technical configurations for permission or allocation. In an iterative manner, different aspects come into play such as: solar irradiance, site
characteristics and infrastructure connection, technology selection and
technical concept, and market and political environment. Every aspect has
a bearing on the feasibility of the whole solar thermal project.
A typical project development approach for CSP does not differ much
in principle from other types of development projects, such as projects to
build conventional fossil power plants, photovoltaic (PV), wind or hydro
power plants or real estate projects. In an ideal world, a pre-feasibility
analysis would be executed for a preselected site and available information
would be gathered. Investing in specialist studies and expensive details
91
© Woodhead Publishing Limited, 2012
92
Concentrating solar power technology
would only be conducted in a limited way, usually with the aim of obtaining
mandatory and essential information.
After a positive outcome of pre-feasibility analysis, a fully-fledged feasibility study would be conducted, supplanting the pre-feasibility initial
assumptions and procuring and elaborating on details. The site qualification
process is then initiated and accomplishes the whole permitting and engineering work, eventually culminating in the signing of the main project
contracts (equipment, O&M (operations and maintenance), environmental) and in financing of the project so that construction can begin.
Through such a careful and step-by-step approach, the main project
development risks will be mitigated properly, fatal flaws can be identified
at an early stage, and substantial amounts of development work is carried
out only on the most promising projects.
CSP, however, has specific features which require a unique approach
based on sufficient knowledge of certain specifics:
•
A project site has to match certain criteria regarding direct normal
irradiance (DNI), land area, topography or slope, water, and interconnection options to grid, road and qualified staff.
• The technical concept and the expected plant performance must provide
sufficient energy to allow an economically feasible plant. Specific care
has to be applied when estimating the plant’s yield, as transient effects
during conversion from fluctuating solar radiation to heat and/or electricity cannot easily be calculated with conventional power software
tools.
• High cost of technology and the resulting high cost of generated power
do not yet supply energy generation at market rates. Incentive mechanisms are required as support which, in turn, set their own guidelines
and create a dependency on political mechanisms and stability.
• High capital costs and the implications of financing costs (interest rates
on debt and return on equity) are the most significant operational
expense during the debt repayment period (typically 15–20 years). Political risk insurance and reliability of energy off-take agreements are
important factors for financing.
The blend of factors that determine the feasibility of a project differs among
projects. Depending on the strategic goals of the promoter or owner, the
applicable weight of selection criteria for each site will be different.
The recommended approach is to start on a general basis, collecting
together all available information to obtain an initial overview of the project
covering all aspects. As the assessment and qualification of the site progresses, the required information will be procured in greater detail, which
will improve the accuracy of the feasibility analysis. This is an iterative
process, where most parts influence each other.
© Woodhead Publishing Limited, 2012
Site selection and feasibility analysis
93
It is beneficial to work from rough draft to detailed analysis as each detail
involves a certain cost, be it the procurement of a DNI measurement station
and measurement data, the optimization of the technical concept or a timeconsuming permissions process, etc. It is important to carry out a fatal flaw
analysis at the beginning of a project to get a clear view of the risks involved
and chances of success.
This chapter aims to give an overview of the process and the challenges
involved. It cannot provide a detailed step-by-step recipe for site assessment, as projects in all locations have strong specific aspects and comparison in most cases is not easy. However, it does aim to provide a guide for
CSP site selection and feasibility analysis.
4.2
Overview of the process of site selection and
feasibility analysis
When it comes to identifying the main criteria for siting or locating a concentrating solar power, or solar thermal energy system, it is preferable to
have a comprehensive checklist and a straightforward approach that perfectly blends into the general project strategy and time schedule.
Site selection and feasibility analysis for a CSP plant is not as simple as
it may appear at first glance. Unlike photovoltaics or wind, where multiples
of identical single units can be installed in parallel and connected on the
electrical side, solar thermal energy does not have a simple system design.
Instead solar thermal energy systems are tailor-made and complete systems,
where the thermal process interconnects the solar field with power generation. In some cases, they provide electricity and combined heat and power,
with the heat used for industrial and/or desalination processes. The application of thermal energy storage and hybrid fuel solutions, be they regenerative or fossil fuels, can also significantly enhance the availability and
dispatchability of solar thermal plants to the point of base load energy
generation.
CSP in 2012 still largely depends on the incentive mechanism provided
at the targeted project location. Examples of this are: grants, government
guarantees, feed-in tariffs, competitive bidding or tendering processes with
government backed power purchase agreements (PPA). The motivations of
governments can change over time, and it can be observed that regulations
for renewable energy projects have been adjusted frequently. Reliability of
regulations and political stability in the long term is an important aspect
when selecting a target market for a project location. Localization of manufacturing and supply of materials is an important aspect for most governments when introducing incentive mechanisms. However, with political risk
insurances for imported items through export credit agencies (ECA) and
loans from developed countries favouring supplies from abroad, this creates
© Woodhead Publishing Limited, 2012
94
Concentrating solar power technology
a general conflict regarding sourcing of materials. An additional aspect that
needs mentioning is that international sourcing is a typical practice in order
to obtain the lowest cost items.
When looking at the timeline for development and implementation of a
‘typical’ project, it usually takes at least 2–3 years from the initiation of
project development until start of construction, with an additional 2–3 years
for construction, plus the expected lifetime of 30+ years for the plant and
a loan period of 15–20 years.
A typical site identification process, if started from scratch for a completely unknown market, is illustrated in Fig. 4.1. The sequence and aspects
of this process are described in the following sections.
4.2.1
Market analysis
The selection of a specific market should be accompanied by an assessment
of the overall market facts and perspectives for CSP. The function of market
analysis is to verify whether the existing basis and outlook of the market
are, in principle, suitable and provide sufficient prospects for feasible CSP
projects. This assessment would typically take into account and analyse the
energy market, off-take options and energy price ranges, and government
support with its political and legal frameworks.
4.2.2
Regional or national study and site identification
When targeting a preselected CSP region or country for siting of a solar
thermal power plant, it is desirable to obtain a comprehensive overview of
the entire area. Instead of the substantial effort of travelling the whole
region to look for prospective sites, a geographic information system (GIS)
can be used to identify preferred regions. ‘A geographic information system
is an information system that is designed to work with data referenced by
spatial or geographic coordinates. In other words, a GIS is both a data base
system with specific capacities for spatially-referenced data, as well as a set
of operations for working with the data’ (Star and Estes, 1990). Therefore
a GIS-based analysis makes use of many different remotely sensed information layers and information obtained from ground-based surveys or a combination of both. The GIS requires an input of the most reliable data sets
which include direct normal irradiance (DNI), topography, road system,
electric grid, surface water and hydrological maps, etc. The GIS, then, based
on pre-defined criteria and a selection algorithm, highlights the best matching regions for a CSP site. The preselected regions are then assessed in more
detail. An example of a software-based selection process with integrated
GIS, covering multiple aspects is described for North Africa in Broesamle
(1999) and Broesamle et al. (2001).
© Woodhead Publishing Limited, 2012
Site selection and feasibility analysis
Market analysis
• General suitability for CSP
• Identify focus market/regions
1
Country/region study
• Region and DNI mapping (GIS)
• Infrastructure analysis
• Pre-selection/securing of project sites
• Target definition of technical concept
2
Pre-feasibility and detailed feasibility analysis
(iterative process - from rough to detailed)
• DNI assessment/measurement
• Site assessment
• Techno-economic optimization
• O&M concept
• Energy yield
• Financial yield
3
4
Project qualification
• Geotechnical and topographic survey
• Environmental impact assessment (EIA)
• Obtain permits and authorizations
• Expert opinion on DNI and yield
• Contract negotiations equipment/EPC, O&M
• Obtain EPC and equipment price quotes
Financial closing
• Due diligence (legal, technical, financial)
• Construction contract
• O&M contract
• Equity and debt agreements
5
4.1 CSP site selection and qualification process (© Martin Schlecht,
Suntrace, 2012).
© Woodhead Publishing Limited, 2012
95
96
Concentrating solar power technology
Eventually, travel to the region of interest is necessary to verify the
GIS findings at the location (ground truth), and with the use of accurate
GIS system data and thorough analysis, these visits can be much more
efficient and less time consuming. Personal visits also allow gathering more
information regarding the neighbourhood, availability of land, contact with
local authorities and persons, and so on. Following these site visits, a ranking
of sites should be made based on individually defined criteria. As an
outcome, the selected location(s) can then be subject to a site-specific
pre-feasibility.
Some regional studies have been conducted through public funds and by
research institutes and government agencies. An example is the analysis of
renewable energy potential for North Africa, as initiated by Broesamle
(1999), which was continued in much more detail and covering broader
aspects. The results are presented in several studies: Med-CSP (Trieb et al.,
2005), Trans-CSP (Trieb et al., 2006) and Aqua-CSP (Trieb et al., 2007).
Another example of country-wide assessments of CSP resources using
GIS is presented by Mehos and Perez (2005). Their analysis of DNI data,
‘combined with geographical information system (GIS) data, has quantified
the solar resource potential for large-scale power generation using CSP
technologies. . . . Prime locations for future solar power plants can also be
identified by factoring in information on constraints on electricity transmission and access to load centers, which are the regions where electricity is
consumed.’
Stoddard et al. (2006) has also applied a GIS-based approach for a study
on CSP potential and benefits in California.
4.2.3
Pre-feasibility analysis
The aim of a pre-feasibility study is to assess at an early stage and on an
indicative but comprehensive basis the general feasibility of a pre-selected
project site. Starting the pre-feasibility analysis is not straightforward as all
aspects influence each other significantly. Pre-feasibility would typically
include a fatal flaw analysis of the project site, a first assessment of site
parameters to decide on potential technology configurations and a complete overview of project economics. The whole approach is cyclical, and
adjustments of a single parameter may impact the overall result requiring
an iterative procedure as shown in Fig. 4.2. This would include assessment
of project site, infrastructure and solar resource. Initial technical concepts
would be developed and modelling of energy yields and financial yields will
give first results.
Initial numbers and figures can usually be based on qualified assumptions
and publicly available data, so that the expenses involved with gaining
precision and detail can be deferred to a later stage when a positive
© Woodhead Publishing Limited, 2012
Site selection and feasibility analysis
EIA
Grid
Social
labour
PPA/
REFIT
Land
Financial
assumptions
Water
Iterative update process
with precise information
DNI
• Detailed boundary conditions
• Firming up cost assumptions
• DNI data (measurements)
• Optimize technical concept
• Update financial assumptions
97
O&M
concept
Collect
input
parameters
Financial yield
(financial model)
Technical
concept
Energy yield
(performance
model)
4.2 Process from pre-feasibility to a bankable project through iteration
with continuously more detailed and realistic data (© Martin Schlecht,
Suntrace, 2011).
conclusion for the project site has already been made. Therefore, the degree
of detail and the accuracy of information may be kept brief on a case-bycase decision.
Sensitivities of main project parameters, such as the range of expected
DNI, equipment prices and off-take prices can give perspectives of worst
case and optimistic scenarios. In case of multiple sites under consideration,
a parallel pre-feasibility analysis can exploit synergies which allows a
ranking of the sites. Such multiple site assessment within a pre-feasibility
analysis is given by Stoddard et al. (2005) as an example regarding selection
of two options for locations with the US state of New Mexico. Since most
projects are pursued by private entities and are not usually accessible in the
public domain, the number of published examples is limited.
4.2.4
Feasibility analysis
The feasibility analysis continues seamlessly from where the pre-feasibility
analysis concludes. The aim of the feasibility analysis is to obtain more
© Woodhead Publishing Limited, 2012
98
Concentrating solar power technology
detailed information on a project that has a high chance of realization, and
by doing so clarify most of its aspects and address most of the concerns
associated with it. An abundance of data, which may have been taken as
qualified assumptions or on the basis of educated guesses will need to be
verified and made more detailed during the course of this analysis. For the
feasibility analysis, specialist studies will be required: a DNI measurement
campaign and solar resource assessment, the use of these measurements in
combination with long-term satellite time-series, geotechnical and topographical site assessment, engineering for technical concept and energy
yield modelling, permit engineering, environmental and socio-economic
assessments, financial modelling, legal guidance on administrative
processes.
Just like during the pre-feasibility stage, the process is not straightforward. The results in one area could influence conclusions drawn from results
in another area to a certain extent. The whole approach remains circular.
Through several rounds of iteration, the conclusions will be refined as
shown in Fig. 4.2. As mentioned above, most projects are conducted by
private entities, and feasibility reports are usually not in the public domain,
and therefore there is a scarcity of published examples.
4.2.5
Project qualification phase
During the project qualification phase, the required work will be performed
to develop the project until it is ready for financing and the start of construction. This includes the process to apply and obtain all required permits
and authorizations, conclude all required project contracts, prepare for
equity and debt funding, and secure the technology. The results from the
feasibility analysis will be used as the starting point for this phase.
The main layout of the plant has to be finalized, as significant parameter
values will be required as input for permit applications (water consumption,
water discharge limits, plant footprint, land area covered and owners
affected, route for transmission line, technical plant and operation patterns,
etc.). At an advanced permitting stage, time management is vital, every
adjustment to the system design will require additional work, involved cost
and additional time to reach approval and receive the required permits.
As the project progresses, the adjustment of figures and details may
require a frequent review and update of the project feasibility. It is expected
that assumptions and results will be further refined, reducing uncertainty
of the overall project economics to a level considered reliable for financing
the project. Towards the end of the project qualification phase, several
independent expert opinions need to be obtained to support the most relevant figures, such as detailed assessment of solar resources and energy
yield. In the project qualification phase, at least one year of measurements
© Woodhead Publishing Limited, 2012
Site selection and feasibility analysis
99
should be completed and, together with updated site-specific satellitederived data, these form the base for site-specific meteorological data sets,
which should be as precise as possible to minimize uncertainty of the potential power output.
4.2.6
Finalization of contracts and start of construction
At this stage, permissions and authorizations, off-take agreements, land
contracts and many other items are completed, and the last project feasibility update has proven the assumptions correct. At this point, a risk assessment and due diligence assessment of the whole project must be carried
out. All the project details will be required for this, and thorough preparation will assist and ease the process. Unresolved issues may appear and may
need resolving under increasing pressure and time constraints. During this
period, all main project contracts are finalized, such as the construction
contract, off-take agreement (PPA), grid interconnection agreement,
land contract, rights of way for transmission line, piping and roads and
O&M contract. Equity and debt agreements shall be arranged and the final
version of the financial model established.
4.3
Main aspects considered during the
pre-feasibility and feasibility phases
The goals and requirements for the solar thermal project need to be clearly
determined in the first instance, taking into account all criteria. Locating
a site for a CSP system usually requires an individual approach, depending
on a blend of fulfilment of prerequisites, the requirements of the system
and existing boundary conditions. Important aspects and criteria for a
site have a specific impact on the technical and economic feasibility of
the project. Each criterion also influences the others, creating a circumferential relationship. An iterative analysis process is therefore required (see
Fig. 4.2).
4.3.1
Economic assumptions
The cost of the project, achievable off-take prices and the terms and conditions for the financing of the project are significant factors in terms of
project feasibility, and should be reviewed initially and on a more general
market approach basis, before investing too much time (and money) on site
studies. A prerequisite for a feasible project is a viable off-take agreement
for the plants’ products (electricity, heat, water, pressurized air and other
elements that are applicable).
© Woodhead Publishing Limited, 2012
100
Concentrating solar power technology
4.3.2
Solar irradiation
There should be an appropriate amount of solar irradiance for the project
to be feasible, but the minimum threshold has to be determined for each
project on an individual basis. DNI converts to energy, which converts into
money and has to provide the expected returns for the project. If the DNI
data is very accurate, this will reduce the uncertainty of the project economics directly, and will influence the base case for the project economics.
Details of solar resource assessment have been discussed in Chapter 3. With
projected reduction in total project costs in the future, lower DNI regions
will become more economically attractive at the same energy off-take
prices.
4.3.3
Land, topography and soil
Siting has to reflect land plot borders related to ownership. The next step
is to collect requirements for slope and soil, which are different for the
various CSP technologies and need to be considered. Levelling and terracing works can be done but both incur extra costs. Soil replacement, earth
movement or extensive foundation works are also determinants of cost.
How much the project can afford has to be determined on an individual
basis. A GIS system with an detailed topographic map included can help
to identify areas with acceptable slope and topography as described in
Broesamle (1999).
4.3.4
Water
Water availability can enable the possible use of wet cooling systems for
the power block, which have a significant economic advantage over dry
cooling systems as they are more efficient and involve a smaller financial
investment. Dry cooling, however, is a feasible option from a technical point
of view, and the project economics may have to support the use of this
technology.
4.3.5
Infrastructure
Grid access
The main issue concerning infrastructure is the connection point to the
electric grid for power evacuation. The distance to the electrical grid is
associated with a cost per km and the upgrade of facilities at the grid tie-in
point.
© Woodhead Publishing Limited, 2012
Site selection and feasibility analysis
101
Interconnection with other plants and processes
If the plant will be supplying products other than electricity, the interconnections with the customers for those other products need to be considered.
For example, if the plant will produce heat, the consumers need to be situated adjacent to the CSP plant.
Roads and highways
The accessibility of the plant is important. During construction, heavy
hauling of the main equipment items and all deliveries of materials and
construction machinery and vehicles must be possible. During operation,
delivery trucks and maintenance must also have access. Therefore costs of
access roads and potential upgrades of bridges for heavy hauling may have
some impact on the cost of the project.
4.3.6
Environmental impact assessment (EIA)
An EIA study is usually required. It has to reflect the applicable local standards and will most likely conclude with certain restrictions on the technical
concept and the plant layout and may define mitigating measures for the
project. Animal and plant habitats and nature conservation and protected
areas must be respected to avoid significant problems and delays during
permitting.
4.3.7
Population and labour
A site in the vicinity of larger villages and/or cities can ensure the availability of qualified local staff both during construction and O&M phases. If
qualified personnel have to be brought in from distant locations, or even
from foreign countries, this also increases the cost.
4.3.8
Socio-economic impact assessment
The various facets of the social impact of the project on the prescribed
region often need assessment, but the depth of assessment can vary significantly. The socio-economic impact assessment usually assesses the sustainability of a project in the host country/region. Involvement of municipalities
in the development process with respect to their role as authorities but also
as a stakeholder is critical. In this case the host country or region will expect
to gain certain benefits in terms of economic development from the project,
such as tax revenues, improvement of infrastructure, qualification and
employment of local persons, etc. The benefits may also include the
© Woodhead Publishing Limited, 2012
102
Concentrating solar power technology
economic development of local businesses such as for manufacturing, maintenance works, etc.
In countries like South Africa, the socio-economic development aspect
is almost institutionalized (with respect to the involvement of historically
disadvantaged groups), while in other markets, this topic may be covered
in a much more general sense and may not require specific action.
4.4
Boundary conditions for a concentrating solar
power (CSP) project
The market and the available incentive mechanisms define the route to
follow for CSP projects. The general type of application, technology selection and technical concept define the overall approach to a project site in
terms of size, infrastructure requirements and meteorological parameters.
The following gives a summary of criteria, which influence other criteria
and the decision to initiate a project, even before commencing the processes
of site selection and feasibility analysis.
4.4.1
Off-take and market
The energy off-take agreements are the central contracts for any energy
project. The reliability of the agreement and the off-taking party are key
criteria, specifically when the financing of the project relies on a nonrecourse or limited recourse scheme, where the lending bank’s assessment
and criteria define the financing of a project. As off-take agreements are
usually part of the incentive mechanisms, an appraisal of these is equally
important. In the case of government-regulated feed-in tariffs, these are
generally reliable, with the political risk remaining the largest obstacle in
the formula. However, the off-take body introduced by the government
should be a reliable and bankable stand-alone player or at least sufficiently
supported by the government.
4.4.2
Incentives and support schemes
In most markets, CSP will need financial support through incentives in the
next few years for the project economics to be feasible, as cost reduction
measures will not lead to grid parity at wholesale level in the near future.
IEA (2010) states ‘In the sunniest countries, CSP can be expected to become
a competitive source of bulk power in peak and intermediate loads by 2020,
and of base-load power by 2025 to 2030’. One possible incentive mechanism
is feed-in-tariff (FIT) schemes. In the cases of other incentive mechanisms
such as capital grants, the reliability of these should be considered with
equal importance. As CSP projects are capital intense, a long-term
© Woodhead Publishing Limited, 2012
Site selection and feasibility analysis
103
investment horizon of 15–18 years is usually required. In cases where
process heat will be provided simultaneously, or used via cogeneration to produce desalinated water, for example, and a blend of
revenue streams provides the economic basis for the project, the matter
becomes progressively more complex, as most incentive schemes currently
do not provide for the remuneration of solar generated heat or other products. The assessment of a project should also examine the possibility of
failure to qualify for an off-take agreement such as a PPA or FIT, if for
example, competitive bidding or tendering is required. It can be of value to
have a long-term perspective of the market, which is not solely dependent
on a single attempt.
4.4.3
Specification of energy products
The general aim of a project has a major influence on the selection of a site
and technical concept. CSP plants can be used to generate electricity and/
or process steam, and/or they may have other uses such as production of
hot water and pressurized air or heating, cooling, desalination, etc. If solar
generated process heat will be produced, the plant must be located in the
vicinity of the customer. Also the load and capacity requirement from the
heat off-taker might determine and define the technical concept of the CSP
plant. If the CSP plant is to be combined with other, possibly fossil fuel-fired
steam heat and power generating systems, the concept has to be evaluated
comprehensively and most criteria for sizing the solar plant will be defined
by the overall concept. In the case of solar electricity generation, the vicinity
and accessibility to the electric grid plays a key role, while giving a broader
range of possible locations for such projects. In this case, apart from grid
and DNI, also water, roads and other infrastructure criteria will become
part of the evaluation. However, usually there is no ‘standard technical
concept’, as most projects will require tailoring of technology to the projectspecific boundary conditions.
4.4.4
Dispatch mode: storage and hybridization
The energy production of a CSP plant that is ‘solar only’ will follow the
availability of sunshine hours with sufficient DNI, starting up on a daily
basis after sunrise, with peak production at sun peak, fluctuating according
to weather patterns and shutting down with sunset. A CSP system can be
enhanced with thermal energy storage and be hybridized with fuel firing
systems (fossil or renewable fuels) increasing the full-load operating hours.
These features will also lead to an increased dispatchability of the plant,
which supports the management and stability of the electricity grid. These
plants could even provide base-load energy or shift energy generation to
© Woodhead Publishing Limited, 2012
104
Concentrating solar power technology
Solar
TES
Back-up firing
Fossil steam generation
CSP – solar only
CSP with 15% back-up firing
CSP with 8 h TES and
15% back-up firing
CSP with 16 h TES and
15% back-up firing
CSP with 16 h TES 15% back-up
firing and fossil steam generation
Solar augmentation of
coal firing power
ISCC Solar augmentation of
gas fired power
0
10
20
30 40 50 60 70
Annual capacity factor (%)
80
90 100
4.3 Achievable capacity factors for different CSP storage and hybrid
concepts (© Martin Schlecht, Suntrace, 2011).
specific times of the day when it is most needed (peak load). Figure 4.3
provides an overview regarding technical concepts and their potential
capacity factors (Schlecht, 2011).
Enhancements with fuel firing can be on an incremental basis from 0%
fuel firing up to almost 100% fuel firing. Booster firing to increase steam
temperatures to in excess of 500°C can increase the thermodynamic efficiency, while supplemental heat transfer fluid (HTF) heaters can mitigate
cloud periods and keep the HTF warm during the winter season and shutdown periods. Implementation of parallel steam generators can also add to
the solar plant’s ability to provide base-load energy. Conventional steam
generating units can be combined with a solar boiler as back-up and also
extend operating hours up to base-load. These plants can be designed to
match load curves of the grid as required by the operator. At the other end
of the spectrum, a plant in which a solar field is attached to a large conventional power plant (coal- or gas-fired) generates only a very small percentage of its energy from solar resources.
4.4.5
Regulatory restrictions or technical plant concepts
Any incentive scheme typically brings a set of rules, which are imposed on
a project in order to qualify for the incentive. This often affects the technical
concept, as certain technical features may be excluded, required or restricted.
Current global feed-in schemes focus on solar generated electricity only,
and no general incentive scheme currently features solar process heat or
combined heat and power (CHP) applications. The implementation of
© Woodhead Publishing Limited, 2012
Site selection and feasibility analysis
105
thermal energy storage (TES) is also not explicitly supported by current
incentive schemes. For example, there are no schemes that remunerate the
higher value of dispatchable electricity. Examples of policy restrictions and
effects on technical concepts include:
•
•
•
•
Spain only permits co-firing by fossil fuel up to 12–15% related to the
annual energy production. If this is not the case the plant cannot qualify
for the feed-in tariff. Additionally, plant capacity is capped at 50 MWe
by law for renewable energy plants. This restriction, however, does allow
the application of TES, which enables CSP plants to generate solar
electricity also during night time on the same remuneration basis. The
Andasol 1 project, which has been operational since 2008, includes a
molten salt-based two-tank thermal energy storage. A total of 28 out of
61 pre-allocated CSP projects in Spain will implement thermal energy
storage with more than six equivalent full-load hours capacity according
to Protermosolar (2011).
The Solar Nevada One project in Nevada, USA, is only allowed to use
a 2% energy contribution from gas for freeze-protection measures. This
project reached an individual agreement for its PPA and governmental
support.
Solar-fossil hybrid CSP plants only comprise a few complete systems
presently, and each of them is based on individual regulations for the
energy off-take. The integrated solar combined cycle (ISCC) projects in
Morocco, Algeria and Egypt have been set up under a specific incentive
funding scheme from the World Bank’s Global Environmental Facility
(GEF).
The United Arab Emirates/Abu Dhabi located Shams 1 project from
Masdar has fossil-fired booster-firing to superheat live steam to 540°C.
The technical concept was determined by a public tender process, which
was focused on lowest electricity price.
It is apparent that each market and incentive scheme requires an individual
approach to integrate the specific regulatory boundary conditions into the
feasibility.
4.4.6
Overall project viability
A general assessment of the solar resources of a target region should be
made as a first step in order to map the distribution of solar irradiation.
DNI directly converts into energy produced and is a major influence on the
income stream from energy sales.
A general conclusion is that DNI values below 2,000 kWh/m2/a may not
yield a viable project, but the levelized cost of energy of a specific project
must show parity with the energy off-take price achievable in the market.
© Woodhead Publishing Limited, 2012
106
Concentrating solar power technology
This price is specific for each market or for a single project in the same
market. Both figures define the investment the project can carry. Eventually
the cost of the project must match the prices of technology and financing.
Every project promoter must assess their individual threshold for DNI in
the specific market or project depending on the desired technology, and
then decide on the detailed project location. As equipment prices are poised
for reduction over the next years, new markets will open up when levelized
cost match off-take prices achievable.
4.4.7
Long-term perspective: political stability
Considering the capital intense investment into a solar thermal plant, a
long-term perspective for the project is essential. This should factor in the
following aspects:
•
•
•
reliability regarding level and duration of off-take agreements
off-taking party/company stability and commitment
political security regarding the long-term financial security of the project,
e.g. no uncompensated expropriation, fulfilment of incentive schemes,
and no retroactive changes to FIT or PPA prices.
Protection against some of these aspects may be possible with commercial
insurance companies or government agencies, such as export credit agencies
like the German Hermes. In the end, this aspect is part of the risk assessment and can usually be mitigated to a large extent. However, as this mitigation is part of the financing strategy of the project, fatal flaws should be
identified in the initial approach to the project.
4.5
Detailed analysis of a qualifying project location
4.5.1
Site-specific solar resources and
meteorological patterns
Direct normal irradiation
The Direct normal irradiance (DNI) is one of the most crucial aspects when
assessing and optimizing the technical concept for a CSP plant. Much
emphasis should be placed on the thorough determination of the relevant
DNI for a particular project and its specific location. Every reduction in
uncertainty in terms of the solar resource will directly result in a better
predictability of the energy production capabilities of a particular concept
(which can be determined through performance modelling).
Microclimate can have a significant influence on the DNI at a specific
location. There is a high chance of introducing an error if analysis results
from a larger area are used instead of assessing the specific DNI of each
© Woodhead Publishing Limited, 2012
Site selection and feasibility analysis
107
potential location. On-site DNI measurement for at least a full year, in
order to cover a complete seasonal cycle, is usually required for a project
at the stage of financing. This data, when properly applied, can be correlated
with overlapping satellite data. This, together with historical satellite data
(stretching back a minimum of 10 years) can be used to project the specific
solar potential of a project site for the lifetime of the plant (Meyer et al.,
2008). Without on-site readings, the uncertainty in terms of the DNI could
be as high as ±30% in some regions such as India or northern Africa.
Wind
Strong wind speeds can have a direct impact on power generation, as they
can affect the focusing of sunbeams by creating vibrations and bending
mirrors. The various technologies and makes have individual requirements
in this regard, so cut-off wind criteria should be requested from applicable
manufacturers. Countermeasures such as wind-breaking barriers and fences
can mitigate this effect. When looking at the annual average, the influences
of wind are in most cases fairly small, and it will not have a critical influence
on the overall economics, as long as it is factored into the calculations.
Ambient conditions
Other meteorological parameters such as ambient pressure, temperatures
and humidity have some influence on the overall plant performance, thus
they must be assessed together with DNI and wind. Wet and dry bulb temperatures define the achievable cold-end temperature of the power plant,
and thus have an influence on the steam turbine design and efficiency. This
is of relevance for both wet and dry cooling systems.
Weather patterns
Clouds, rain and other weather patterns are included in the DNI assessment, as these factors have direct influence on DNI. A long-term assessment
based on satellite data will provide the best estimate for the plant operational lifetime. If weather patterns indicate strong rain, stormwater run-off
and flooding concepts should be assessed individually taking into account
the topography of the site and surrounding area.
4.5.2
Land and surroundings
Orientation and slope
Generally a north–south orientation of the land plot is important. As all
technologies can cope with some degree of slope, the gradient should
© Woodhead Publishing Limited, 2012
108
Concentrating solar power technology
preferably face towards the equator. Parabolic trough technology accepts
less than 2% slope while linear Fresnel reflector systems can accommodate
up to 5% slope. Solar tower systems can accommodate a steeper slope, as
long as access for construction and maintenance is sufficient and the slope
supports the arrangement of the heliostat field. The feasibility of building
on steeper slopes also depends on the tower technology concept, which
varies considerably between different designs. Dish systems can be installed
as distributed arrays; so the only relevant factor to consider would be units
shading others, so long as access was acceptable.
For parabolic trough systems, the solar field is usually divided into several
subfields, which could allow a terracing approach with different elevations
for each subfield. Again, this would mean that a site with an initial slope
steeper than 2% could be used; however, such terracing works are usually
quite cost intensive, so this cost must be considered in the overall economic
evaluation.
Topography and soil
Should the slope not be perfect and continuous over the requirements for
area, earth and levelling works must be assessed. This should also include
an analysis of the soil conditions, as these are a most crucial input for sizing
and dimensioning of foundations and civil works design. Substantial subsoil structures might be required on sites with soft soil, while sandy topsoil
might need to be replaced or removed. In turn a rocky surface might need
severe treatment when it comes to levelling works. Soil patterns might vary
across a typically large project site of several square kilometres, so a thorough soil analysis over the whole area should be part of the assessment
procedure.
Free horizon
No obstacles in the vicinity should shade the mirrors, preferable down to a
minimum angle (e.g. 3°) above the horizon. Given the anticipated 20–40
year lifetime of a plant, it should be checked that no future developments
in the vicinity are likely to create such obstacles. This assessment should
also take into account the fact that dust and aerosols (e.g. vapour from
cooling towers, smoke-stacks, etc.) can reduce the DNI partially.
Footprint and scaling
The footprint of a solar thermal energy plant is scalable with the installed
capacity. However, some restrictions have to be considered from the technical side:
© Woodhead Publishing Limited, 2012
Site selection and feasibility analysis
109
•
Parabolic trough (PT) and linear Fresnel systems typically apply certain
modular length patterns to one particular loop, which then would be
applied in serial/parallel arrangements. Most PT plants today require
about 40 × 300 m for one single loop, with 100+ loops for a simple
50 MWe plant in Spain, while a linear Fresnel loop could require a
straight 25 × 1000 m for one loop, with 22+ loops for a simple 50 MWe
plant in Spain. New developments on PT aim at larger aperture width.
• Solar tower plants are preferably built on circular or semi-circular
shaped footprints, but also rectangular fields are feasible.
• Additional space for service roads surrounding the solar field, fencing,
possible wind-breaking measures (earth banks, walls or vegetation)
should be considered at the boundary of the plant.
• Depending on the outcome of the environmental impact assessments,
specific requirements could be imposed on a plant, which require additional efforts.
With these patterns, one can see how a plant layout could fit into a given
land plot and its usually non-technical provided borders. However, with a
preselected technology the requirements for a solar project site can be more
specifically defined.
Ownership structures
The legal situation of the required property is also a crucial selection criterion. When approaching the legal side of ownership structures the following
questions should be considered: Is the land public or private? Will it be a
direct purchase or a purchase option or a lease option? How secure is the
land for the project throughout a possibly lengthy development process?
Also depending on the type of financing, the required legal certainty on the
complete property becomes a crucial matter, as agreements with all owners
have to be completed and the surface rights for the installation of the CSP
system should be recorded in the title of the land.
4.5.3
Infrastructure interconnections
In addition to a suitable land area with sufficient solar resource, the standard of the surrounding infrastructure is of significant importance.
Electricity grid
The generated power must be evacuated from the plant; therefore it needs
to tie into a high voltage (HV) electricity line with sufficient voltage level.
Projects of more than 20 MW capacity usually have to look for lines of
60–400 kV, depending on the applicable grid code and network voltage levels.
© Woodhead Publishing Limited, 2012
110
Concentrating solar power technology
Determination of a feasible tie-in point always requires the cooperation
of the grid operator, who needs to analyse the capacity of the proposed
tie-in location in relation to the capacity planning of the whole network.
Tie-in into a HV line by building a new substation may be an option for
larger plants, but most likely one has to tie in at an existing substation, which
also needs to have the available capacity. The procedures for getting a
capacity reservation for a project are different in each country, so one has
to adjust to the local codes and requirements.
Distance to the grid should be kept minimal, as each km of transmission
line requires additional investment. Larger projects usually can afford
longer transmission lines, as the specific cost for the transmission line can
be utilized by a larger amount of energy.
The routing of the transmission line has to be approached carefully, as
usually permits and rights of way related to the properties along the projected route need to be secured upfront, and reaching consent with the
respective authorities and landowners can involve a time-consuming effort.
Also environmental concerns must be considered, as a specific environmental impact assessment for the transmission line is required in most
countries.
Road network
The connection to the road network serves two purposes. During construction of the plant, all materials have to be transported to the project site by
trucks. As this includes heavy equipment, the route for heavy hauling (e.g.
the capacity of road bridges from ports of landing for overseas shipments)
has to be considered. Some improvements required specifically for construction transport can be on a temporary basis.
For operation of the plant, access roads should be erected as permanent
structures. They should suffice for all regular trucking of materials and
during maintenance periods. It should be considered that, during the 20–40
year operation period, major overhauls of steam turbines and generators
would be made, in addition to possible future improvements to the installation. It will be a strategic decision between actual needs and benefits for
further requirements. The cost and authorization effort for this road will
also depend on the distance to the existing public road network.
Fuel availability
For most CSP plants, some small degree of supplemental fossil fuel firing
is considered. Fuel availability and specific cost for transporting it to the
project should be considered. For larger fossil fuel shares, such as an ISCC
(integrated solar combined cycle) system, the fuel sourcing becomes a much
more significant aspect for the feasibility of the project. It may mandate, as
© Woodhead Publishing Limited, 2012
Site selection and feasibility analysis
111
additional location criteria, the vicinity of a pipeline, or a train connection
for liquid or solid fuels.
4.5.4
Hybridization with other fuels
A hybrid concept with a reduced solar share may lead to a complete shift
in perspective compared to a fully stand-alone solar electric generating
facility. The availability of fuel to the plant becomes more important with
the higher degree of fuel compared to solar resource input. The vicinity to
fuel infrastructure (pipelines, ports, railways) may become a primary criterion for site selection. If a solar field is being added to an existing conventional energy plant (coal, gas, industry application), the siting of the plant
has to be in the vicinity of the existing plant. The process boils down to the
general applicability of CSP, the technical concept and the commercial
feasibility when overcoming whatever compromise might be required with
given boundary conditions.
4.5.5
Water: sources, uses and related requirements
When it comes to the water consumption during operation, the discussion
focuses very much on the cooling system for the steam turbine condenser. As
the highest DNI usually appears in desert areas, typically regions with very
little water resource (aside from salt water at sea shoreline), the usage of
large volumes of water may be restricted or it may simply not be available.
Water qualifying for cooling purposes could be any surface water – salt
and fresh water both can be applied – or subsurface water (from wells). If
wet cooling is an option, hydrology and groundwater availability and accessibility should be examined in detail. The biggest obstacle will most likely
be the authorization for water use, so the permitting and legal processes for
this should be part of the feasibility analysis for a project.
Dry vs wet cooling technologies
From an economic perspective, wet cooling (usually with wet cooling
towers) will always be the preferred method, as long as water is available
and affordable. Dry cooling is technically feasible for all CSP technologies,
and is not a technology risk, as the technology has been implemented in
conventional power plants over the globe for a long time. The issue with
dry cooling is its negative impact on project economics:
•
Air as a cooling medium has a lower heat transfer coefficient than water.
If air is used, advantage also cannot be taken of the chilling effect produced by evaporation when wet cooling towers are used. Therefore the
exhaust temperature of the steam from the turbine is several degrees
higher. This results in reduced efficiency of the water-steam cycle.
© Woodhead Publishing Limited, 2012
112
•
•
Concentrating solar power technology
Air-cooled condensers require a larger mass flow of air for cooling, so
the fan power required by a dry cooling tower is higher than that
required by a wet cooling tower. This results in increased electricity
consumption within the power plant (parasitic consumption) and
reduces the revenue through power sales.
The investment cost and required land area for air-cooled condensers
are usually higher, so the total investment cost of the project will be
increased.
In comparison to these disadvantages, the benefit of a reduction in water
consumption with dry cooling is in the range of 85–90%. In conclusion, dry
cooling imposes some economic burdens on solar thermal power plants
compared to wet cooling, but significantly reduces the amount of water
required for plant operation and is not difficult to implement from a technical point of view. However, the possibility of lower electricity prices often
gives wet-cooled projects an advantage when competing with dry-cooled
projects during a tender process.
Water requirements
A solar plant requires a certain volume of water. The uses can be divided into:
• 85–90%: Wet cooling tower (evaporation replacement and blow-down/
make-up)
• 15–10%: Process, service, demineralized and mirror washing water.
With dry cooling, even a truck-based supply of water might be feasible.
Water-steam cycle
Demineralized water is required for the water-steam cycle due to the nature
of steam production. Steam generation for subsequent power generation
dominates in all large-scale CSP technologies. Small-scale technologies,
though, such as Stirling dishes or small solar tower concepts use air/gas as
the HTF. Demineralized water should be processed on-site from raw water
in a specific water treatment plant. The volume of water required for this
purpose is not very large. It has very specific purity requirements, though,
which are defined by steam turbine makers. Consumption is usually defined
by blow-down and leakage replacement, except for some specific uses for
shutdown and start-up of the plant.
Process and service water
A power plant usually requires process water for various purposes: cooling
of rotating equipment bearings (pumps, motors, etc.), washing and cleaning
of the plant, etc.
© Woodhead Publishing Limited, 2012
Site selection and feasibility analysis
113
Mirror cleaning
Mirrors need frequent cleaning to maintain proper reflectivity. The cleaning
frequency is specific for each project. It depends on the dust load at the
specific site, and on the economic feasibility of cleaning: cost of water and
labour for cleaning versus increased energy sales through higher reflectivity
of mirrors. For cleaning of mirrors, it is recommended to use demineralized
water to avoid staining the mirrors.
Condenser cooling system
Depending on the type of cooling systems, different volumes of water are
required.
•
•
•
•
Dry cooled condensers will not require almost any additional water
apart from a potentially increased condenser hotwell volume, which
does not increase consumption.
Hybrid cooling towers are a mixture of dry cooling and wet cooling, as
some water is sprayed on the exterior of the condenser tubes which will
evaporate and lower the achievable condensate temperature. Water
consumption will still be significantly lower compared to a wet cooling
system.
Wet cooling tower systems will utilize with roughly 10 times the volume
of water (when compared to dry cooled condensers) during operation
as a consequence of the evaporative cooling in the cooling tower. In
addition, as a consequence of the evaporation process, the constituents
such as minerals and other solids and dissolved solids in the cooling
water will remain and their concentration will increase. The concentration is usually limited (by technical and/or regulatory requirements).
Therefore a fraction of the water needs to be blown down and replaced
with fresh water on a regular basis.
Once-through water cooling could be an option if plants are installed
in the vicinity of water bodies, however this cooling method is not
expected to be a serious option since regions for CSP projects are
typically characterized through water scarcity. This cooling method
requires by far the largest amount of water, even though water is only
heated by a couple of degrees Celsius and then routed back to the water
body.
Water quality and volume requirements
The quality of the available water is important for the determination of
required water volumes and for the design of water treatment facilities. This
is of significant importance at an early project development stage, as usually
© Woodhead Publishing Limited, 2012
114
Concentrating solar power technology
authorizations are required for both water extraction and water discharge.
Environmental regulations will limit volume in supply and discharge, and
will control concentration limits of certain constituents in wastewater
discharge.
The water volume required for the plant operation is based on several
factors:
•
•
•
Projected operating hours and the related heat to be discharged from
the condenser via the cooling system:
䊊
a higher power block efficiency requires a lower specific cooling
water volume
䊊
annual variability of sunshine hours must also be considered, as
during years of more sunshine, operating hours could be up to 10%
higher than assumed during the typical meteorological year (TMY);
this will need to be defined through performance calculations for the
plant.
The environmental permit defining concentration limits for some constituents in the water discharge from the cooling system.
The quality of the raw water to the plant determines the possible number
of cycles in the cooling system, and defines the blow-down rate required
for water discharge and the make-up water volume. Surface water
quality has seasonal cycles, as sediments and concentrations in the water
can be washed into the water from rain or melt water, or also from
industrial/sewage discharges into rivers or lakes upstream of the water
extraction point. A thorough analysis with frequent sampling and laboratory analysis will provide the basis for the design.
In conclusion, the quality of water from available raw water sources is a
significant factor in determining the water volume required for a solar
thermal power plant, in particular when wet cooling is to be applied. The
assessment of water quality and the permissions for water uses and discharge needs special attention, as permits with too stringent requirements
can restrict the operation of the plant after certain volumes are consumed
or if limits are exceeded frequently during regular operation.
4.5.6
Natural hazards risks and mitigation
The risk of natural hazards for the selected area should be assessed, based
on historical events and regional specifics. This assessment should include
the risk of occurrence of earthquakes, tsunamis, bush-fires, flooding from
nearby rivers, severe storm events (e.g. hurricanes, typhoons, hail- and thunderstorms) and volcanic eruptions, as applicable. Mitigation measures
should be developed based on the risk assessment for all risks classified as
relevant.
© Woodhead Publishing Limited, 2012
Site selection and feasibility analysis
115
Part of the risk mitigation should be through technical measures in design
and construction of the plant, while the remaining risk could be covered by
insurance. If a specific natural hazard risk for the selected region dominates,
the insurability against such incident should be confirmed. Otherwise the
financial risk will remain completely with the owner and the financing banks
and, if such risk cannot be accepted, the project location may be considered
unfeasible.
4.5.7
Labour
An important aspect for a project is the availability of qualified staff during
construction and operation of a plant. Construction of a solar thermal
power plant is expected to take between 18 and 36 months depending on
the size of the installation and the experience of the contractor. In this
period several hundred workers will be required for the construction, and
at peak times a multiple of this (up to 2,000 or more workers at a time). All
these workers need at least basic skills, as the quality of the installation
should be appropriate for the expected lifetime of the plant. Higher degrees
of expertise are required for supervisors and construction engineers.
Sourcing of this labour from the vicinity of the construction is advantageous for the owner. Local labour will increase the acceptance of the project
in the municipality and with the population. Training in particular skills and
education of workers for the construction project benefit many households
in the long term, regardless of whether the plant offers equivalent employment in the medium and long term. To bring in all workers from far away,
or from abroad, will be expensive and imposes a logistical challenge for the
contractor.
From the developers’ and owners’ perspective, a fair collaboration with
the surrounding municipalities will be beneficial to the project in the short
and long term. However, balancing the relationship with local politicians
can be a permanent challenge, even during operation of the plant.
Solar thermal power plants require permanent staffing, often on a 24-hour
shift basis. The staff will include experienced engineers, technically skilled
persons and supporting staff (security, housekeeping, mirror-washing, gardening, etc.). A motivated and skilled operations team is a key element for
a successful plant. Preventive maintenance and pro-active optimization of
operations will be of great importance in meeting or exceeding energy
production projections. This is already the case for conventional power
plants, but with the fluctuating solar resource, optimized operation is essential to capture as much of the limited solar resource as possible and convert
it into energy as efficiently as possible.
In remote locations, staff salary packages for qualified staff will be more
expensive. Offers will also need to remain attractive in order to avoid a high
rate of turnover in key positions. The plant commissioning process is a
© Woodhead Publishing Limited, 2012
116
Concentrating solar power technology
unique and important period for the key staff members in terms of their
training to operate the plant, as the strong and weak aspects of the plant
can be observed. ‘Hands-on’ experience will then help to increase their
understanding of the operational behaviour of the plant.
4.5.8
Permissions
Another essential part of a successful project is to obtain all required
permits and authorizations. The regulations to be considered concern every
aspect of the project. The main types of permits can be classified into:
•
•
•
•
•
•
•
•
•
•
•
•
land: register surface rights or full ownership in title of land
tariff/PPA/off-take agreement
grid access and capacity reservation, transmission line rights of way
permits to construct and operate
environmental approval
water extraction and discharge permit, rights of way for pipe routes
road access, rights of way for access road
use of land for industrial purpose (rezoning)
municipal construction permits (civil related)
social and economic impact assessment/economic development
criteria
archaeological clearance
other.
It should be expected that obtaining these permits is a requirement in
most countries. They will be classified differently across the world, though,
and will not always be the responsibility of the same office holders. In some
cases, the allocation of a project is the first step, and following this, obtaining
most other permits is a time-consuming formality. In other cases, each single
permit has to be secured before the project is qualified to receive a feed-in
tariff.
Tender schemes are becoming more popular and are posing a middle way
option, as in most cases project sites and preliminary clearances are prerequisites for successful allocation. It is to be expected that the market and the
incentive schemes will continue to evolve in the future.
4.6
Summary and future trends
4.6.1
Summary
Site selection and feasibility analysis of projects are always integrated into
an overall project development strategy, regardless of whether the project
originated in the public or private sector. A sound and successful site
© Woodhead Publishing Limited, 2012
Site selection and feasibility analysis
117
selection process, in which the guiding criteria and rules are identified in
the beginning, and their impacts on possible technical concepts in a target
region and the resulting cost of energy are analysed, is an integral part of
the development process. The selection of a project site is then the outcome
of the first complete iterative cycle of the feasibility analysis. The term
‘feasibility analysis’ should not be adhered to too strictly, as from the
authors’ point of view, feasibility analysis is an ongoing, iterative process in
which conclusions for a particular project are refined along the development process, eventually concluding in the financial viability of a project at
financial close. If pursued consistently and in a structured way, risks can be
identified and mitigated at an early stage, and sunk costs in the project
development phase can be avoided to a greater extent.
4.6.2
Future trends
The development of sites for CSP projects is of strategic importance in the
context of efforts to increase global implementation of CSP. Apart from
technology solutions (industry) and incentive mechanisms (governments),
the availability of qualifying project sites is a prerequisite for project implementation and the site qualification process usually takes a couple of years.
In markets with clear guidelines, such as Spain from 2006 to 2009, many
project sites have been developed. In other markets where incentive
schemes are not attractive, though, the development of project sites depends
purely on the initiative of the project developer and requires an appetite
for fairly high risks.
In terms of overall global politics and aspirations, it can be observed that
due to climate change and with growing mass production and large-scale
implementation of renewable energy technologies (mainly wind and photovoltaic), the acceptance of renewable energy as a contributor to national
energy supply has substantially improved. Today, even countries like China,
India and South Africa, which still oppose concrete commitments to act
against climate change, are commencing large programmes for installation
of renewable energy capacity.
Given the volatility of the resource (such as wind or sun), renewable
energy brings power fluctuation to the grid system, which will be a growing
challenge for grid operating companies with increasing share of renewable
energy. NREL has assessed this for the US market and came to the conclusion that, as a key difference from PV, CSP when using high-efficiency
thermal energy storage (TES) can be considered a partially dispatchable
resource (Denholm and Mehos, 2011). Due to the storage capabilities, solar
energy can be shifted to peak demand periods, providing firm power, creating additional value and reducing grid integration challenges. The NREL
study concluded that a major benefit of the inclusion of CSP in the energy
© Woodhead Publishing Limited, 2012
118
Concentrating solar power technology
mix is its ability to enable greater penetration of PV (and e.g. wind) than
would be possible without CSP integration. Under such a view, PV and CSP
are partially complementary.
Based on this assessment and a general view of other markets apart from
the US, it is expected that CSP will be a significant contributor to renewable
energy generation, and a mandatory part of the renewable energy mix, in
locations that are at least within reach of solar-rich regions. Even Europe
is considering the import of solar energy from Middle East and northern
Africa (MENA) regions on a large scale.
As renewable energy has to be harvested at source (be it coastal/offshore areas or deserts) and thus has to be transported to the load centres,
the establishment of feasible CSP project locations is part of the global
challenge to add renewable energy to the electric distribution system. Both
central and distributed renewable energy generation will add to the picture,
requiring new grid systems, which are able to shift energy around the continents with high flexibility. Business models for a coordinated trans-national
grid upgrade seem to be lacking to some extent. In particular government
owned monopoly-style utilities tend to resist the refurbishment of the
system instead of moving forward.
With regard to CSP project locations, areas favourable in terms of DNI
will need to be linked to the electricity system. Energy and infrastructure
planning at government level is required to enable these system upgrades.
For example, when looking at Europe and the northern Africa region, a
political effort between the European and African countries is required to
provide the framework for economic agreements and investments into such
infrastructure. Both on a government and industry level, organizations have
been formed to pursue this task, but outcomes are not yet conclusive.
Solar park concepts are coming into fashion. Under such schemes a government initiated solar park authority provides infrastructure such as grid
access, land, water, centralized maintenance, so that multiple solar power
plants can settle in the park area and share these facilities. The success of
such models is still to be observed, but if governments are willing to back
these parks – with regard to off-take of energy, simplified authorization,
clear economic and technical conditions – then this could become an interesting solution.
4.7
References
Broesamle, H. (1999), ‘Solar thermal stations, localization and assessment of the
potential with the planning tool STEPS’, University of Vechta, Doctoral Thesis,
in German.
Broesamle, H., Mannstein, H., Schillings, C. and Trieb, F. (2001), ‘Assessment of solar
electricity potentials in north Africa based on satellite data and a geographic
information system’, Solar Energy, 70 (1), 1–12.
© Woodhead Publishing Limited, 2012
Site selection and feasibility analysis
119
Denholm, P. and Mehos, M. (2011), ‘Enabling greater penetration of solar power via
the use of CSP with thermal energy storage’, NREL, Golden, CO.
IEA (International Energy Agency) (2010), ‘Technology roadmap concentrating
solar power’, OECD/IEA, Paris.
Mehos, M. and Perez, R. (2005), ‘Mining for solar resources, US Southwest provides
vast potential’, Imaging Notes, 20 (3), 12–15.
Meyer, R., Torres-Butrón, J., Marquardt, G., Schwandt, M., Geuder, N., Hoyer-Klick,
C., Lorenz, E., Hammer, A. and Beyer, H.G. (2008), ‘Combining solar irradiance
measurements and various satellite derived products to a site-specific best estimate’, SolarPACES Symp., Las Vegas, NV, March.
Protermosolar (2011), The Spanish Solar Thermal Industry Association (http://www.
protermosolar.com/mapa.html), map with status of projects in Spain (frequently
updated), at December 2011, in Spanish.
Schlecht, M. (2011), ‘Storage and hybridization options’, Presentation at CSP
Summit Seville, Spain, November.
Star, J. and Estes, J. (1990), Geographic Information Systems – An Introduction,
Prentice-Hall, Englewood Cliffs, NJ.
Stoddard, L., Owens, B., Morse, F. and Kearney, D. (2005), ‘New Mexico concentrating solar plant feasibility study’, Final Report to the New Mexico Department of
Energy, Minerals, and Natural Resources, February.
Stoddard, L., Abiecunas, J. and O’Connell, R. (Black & Veatch) (2006), ‘Economic,
energy, and environmental benefits of concentrating solar power in California’,
Subcontract Report NREL/SR-550-39291, April.
Trieb, F., Schillings, C., Kronshage, S., Viebahn, P., May, N., Paul, C., Klann, U., Kabariti, M., Bennouna, A., Nokraschy, H., Hassan, S., Georgy Yussef, L., Hasni, T.,
Bassam, N. and Satoguina, H. (2005), ‘Concentrating solar power for the Mediterranean region’, German Aerospace Center (DLR), Final Report of the MED-CSP
Study for the German Ministry of Environment, Nature Conervation and Nuclear
Safety, April.
Trieb, F., Schillings, C., Kronshage, S., Viebahn, P., May, N., Paul, C., Klann, U., Kabariti, M., Bennouna, A., Nokraschy, H., Hassan, S., Georgy Yussef, L., Hasni, T.,
Bassam, N. and Satoguina, H. (2006), ‘Trans-Mediterranean interconnection for
concentrating solar power’, German Aerospace Center (DLR), Final report of
the TRANS-CSP Study for the German Ministry of Environment, Nature Conservation and Nuclear Safety, June.
Trieb, F., Gehrung, J., Viebahn, P., Schillings, C., Hoyer, C., Kabariti, M., Altowaie,
H., Sufian,T, Alnaser, W., Bennouna, A., El-Bassam, N., Kern, J., Nokraschy, H.,
Knies, G., Möller, U., Aliewi, A. Shaheen, H., Elhasair, I. and Glade, H. (2007),
‘Concentrating solar power for seawater desalination’, German Aerospace Center
(DLR), Final Report of the AQUA-CSP Study for the German Ministry of Environment, Nature Conservation and Nuclear Safety.
© Woodhead Publishing Limited, 2012
5
Socio-economic and environmental
assessment of concentrating solar power
(CSP) systems
N. C A L D É S and Y. L E C H Ó N, C I E M AT –
Plataforma Solar de Almería, Spain
Abstract: A general introduction to the main environmental and
socio-economic aspects associated with concentrating solar power (CSP)
systems is presented. The chapter then analyses the state of the art in the
methods available to quantify the main environmental impacts of CSP
systems. The results of life cycle assessments (LCA) and environmental
externalities assessments of CSP systems are provided. The chapter then
describes the main socio-economic impacts that can arise from the
implementation of a CSP system and provides results obtained using the
input–output method that show increased demand for goods and
services and employment.
Key words: life cycle assessment, externalities assessment, input–output
analysis.
5.1
Introduction
Various environmental and socio-economic drivers are likely to accelerate
the deployment of concentrated solar power (CSP) technologies in the near
future. Consequently the careful assessment of the environmental and
socio-economic impacts of current and future CSP technologies will play a
key role in determining their development pathway. Besides differing in
production costs, the various energy technologies have different collateral
effects on society and the environment. When such effects are not incorporated in the market price of the energy products they generate, they are
named externalities. One of the consequences of the presence of externalities in the energy market is that the resulting energy mix is inefficient from
a social welfare point of view.
Energy market externalities are of various types and magnitudes.
However, compared with fossil fuel technologies, most environmental and
socio-economic externalities associated with renewable energy technologies are positive, resulting in a better social welfare. For example, most
renewable energy technologies contribute to a reduction in emissions of
greenhouse gases (GHG) as well as other pollutants, help diversify and
guarantee the energy supply and are good instruments for the generation
120
© Woodhead Publishing Limited, 2012
Socio-economic and environmental assessment
121
Table 5.1 Energy policy objectives to which CSP systems can contribute
Energy policy objectives
CSP contribution in meeting such
objectives
Guaranteeing economically viable
electricity prices
Guaranteeing security of supply
Climate protection
Very high potential worldwide
Aiming at conflict neutral technologies
Increasing demand for local added value
and labour
Potential for technology exports
Preferring non-intermittent electricity
suppliers
Strong pushing driver
Strong pushing driver
Strong pushing driver
Strong pushing driver
Pushing driver
Pushing driver
Pushing driver
Strong pushing driver
Source: adapted from Viebahn et al. (2008).
of wealth and employment in rural areas, thus contributing to socioeconomic development. Consequently, though most renewable energy
technologies are not yet price competitive in the energy market, their competitiveness is substantially improved when, besides the private electricity
production costs, their associated externalities are taken into account. In
order to do so, and in order to guarantee a sustainable energy mix which
maximizes social welfare, it is important that public decision makers use
economic instruments to incorporate externalities into market prices; in
other words, that they conduct careful assessments to quantify the externalities and assign them a monetary value.
Besides the need to internalize the environmental and socio-economic
externalities, there are other arguments that justify conducting a socioeconomic and environmental assessment of these promising technologies.
According to various experts, there exist various energy policy objectives
which are likely to be the guidelines for the development of the energy
system over the next five decades and, as shown in Table 5.1, CSP technologies can contribute to meeting some of these objectives (Viebahn et al.,
2008). Discussing each of these in turn:
•
•
Guaranteeing future economically viable prices: In the context of ongoing
increases in fossil fuel prices, the extensive exploitation of renewable
energy sources such as solar energy is key to achieving a long-term
decoupling from the fossil energy prices.
Guaranteeing security of supply: By replacing fossil fuel technologies
with renewable technologies, such as CSP systems, it is possible to (i)
increase the reliability of the electricity supply by diversifying the energy
mix and (ii) decrease dependency on fossil fuels.
© Woodhead Publishing Limited, 2012
122
•
•
•
•
•
•
Concentrating solar power technology
Contribute to climate protection: The installation of CSP may reduce
global warming emissions if generating power from CSP plants offsets
generation from fossil-fuelled plants.
Take advantage of energy potentials worldwide: CSP technologies take
advantage of an inexhaustible resource which is very abundant in
various developing countries around the world, where the domestic
technical potentials exceed possible demands by orders of magnitude.
Consequently, there is an argument for the expansion of CSP plants on
the global scale.
Aim towards conflict neutral technologies: Most fossil fuel reserves are
located in geopolitically unstable countries which exacerbates military
conflicts around the world. Another world security threat arises
from the proliferation of nuclear weapons. In this sense, CSP is a more
conflict-resistant technology since it does not involve conflict-relevant
materials. Most importantly, though, the solar resource is abundant and
inexhaustible and thus will not incite conflicts over rights of use.
Increasing demand for local added value and local labour: CSP technology investment decisions have the potential to have a large impact in
terms of local added value as well as employment generation and accumulation of local expertise. Compared to other energy technologies,
CSP technology investments do not require very ‘high-tech’ components, while they do require large amounts of steel, concrete, mirrors
and labour.
Potential for technology exports from countries that are leaders in
technology development: The CSP technology industry in countries that
are leaders in technology development (which includes small- and
medium-sized component suppliers, engineering consultants and large
power companies) has the opportunity to expand its global export
volumes.
Preference for non-intermittent electricity suppliers: Compared to other
alternative renewable energy technologies, by incorporating thermal
storage and co-firing options, CSP technologies are able to offer fully
dispatchable energy at a competitive price level.
5.2
Environmental assessment of concentrating solar
power (CSP) systems
A key benefit of the use of CSP plants is the potential to reduce conventional and greenhouse gas emissions caused by electricity generation. The
installation of CSP may reduce atmospheric emissions if generating power
from CSP plants offsets generation from fossil-fuelled plants. In order to
estimate these benefits several methodologies can be used. In this section
results will be provided for CSP systems from two well-recognized meth-
© Woodhead Publishing Limited, 2012
Socio-economic and environmental assessment
123
odologies and compared with results for other energy generation technologies. The two methodologies in question are:
•
•
life cycle assessment
environmental externalities assessment.
Apart from the beneficial effects described above, the deployment of CSP
can cause some unintended environmental impacts. The main impacts are
impacts on amenities and relate to the large area required for the technology. The main impacts identified are the following (IEA, 1998):
•
•
•
•
visual impacts
noise
ecological impacts due to land use
water resources impacts.
Most of these impacts are local and are therefore highly affected by the
siting of the technology; some of them can be minimized by a sensitive siting
choice. CSP plants using conventional steam turbines to generate electricity
have a requirement for condenser cooling which has until now been satisfied using evaporative cooling towers consuming fresh water. Because solar
abundance and fresh water constraints often coincide geographically, the
cumulative impacts of installing numerous CSP plants in a region raise
policy concerns. The trend is towards more freshwater-efficient cooling
technologies (Carter and Campbell, 2009).
5.2.1
Life cycle assessment of CSP systems
Life cycle assessment (LCA) is a method for systematic analysis of environmental performance from a cradle to grave perspective. This analytic
tool systematically describes and assesses all flows that enter into the
studied systems from nature and all those flows that go out from the systems
to nature, all over the life cycle.
The interest in LCA started in the 1990s and since then a strong development has occurred. The practice of LCA is regulated by the international
standards ISO 14040 and 14044 (ISO, 2006a,b), and there are several introductions (Guinée et al., 2002; JRC IES, 2010) and databases (Ecoinvent,
2007) available. LCA is a robust and mature methodology, although some
aspects are still under development. A thorough review of the recent
advances of the methodology can be found in Finnveden et al. (2009).
A complete LCA study consists of four steps:
1. Definition of the goal and scope of the study.
2. Life cycle inventory (LCI phase) where the collection of all the environmental inflows and outflows takes place.
© Woodhead Publishing Limited, 2012
124
Concentrating solar power technology
3. Life cycle impact assessment (LCIA) phase.
4. Interpretation of the results.
However, it is quite typical that some LCAs only perform the inventory
analysis, delivering a list of emissions, or only evaluate some of the impacts
(like global warming impacts).
There exist several environmental assessments of solar thermal technologies in the scientific literature. Lenzen (1999) evaluated the GHG emissions
of different configurations of CSP plants hypothetically located in Australia.
Weinrebe et al. (1998) performed a life cycle assessment of two plants, an
80 MW SEGS (solar energy generating systems) plant and a 30 MW
Phoebus power tower. Viebahn, within the SOKRATES (Viebahn, 2003)
and INDITEP (Viebahn, 2004) projects, also conducted LCAs of different
configurations of solar thermal plants, a direct steam generation (DSG)
plant, a SEGS plant, and a FRESNEL-type plant. Pehnt (2006) conducted
a dynamic LCA showing the evolution of the impacts of several renewable
technologies including CSP. Lechón et al. (2008) performed an LCA of two
CSP plants in Spain. Within the EU project NEEDS (www.needs-project.
org) an LCA of current and future configurations of CSP plants was performed showing also the evolution of the environmental performance of
this technology with time.
All of these studies show important benefits in terms of reduced environmental impacts for solar thermal power plants compared to other competing electricity generation technologies, especially in terms of emissions of
greenhouse gases. Burkhardt et al. (2010) performed an LCA of a reference
design of a parabolic trough CSP facility in California and evaluated the
effects on LCA results and freshwater requirements of different power
plant designs. GHG results of these studies are summarized in Table 5.2.
Values of global warming emissions in solar only operation reported
in the literature range from 11 g/kWh to 60 g/kWh and from 12 g/kWh to
90 g/kWh for central tower and parabolic troughs, respectively. These emissions are well below the emissions produced by conventional electricity
generation sources (see Fig. 5.1). Discrepancies among studies can be due
to several factors including methodological issues such as the GHG emission intensity of the different materials, technological differences in the
plant configuration, as well as some issues related to the location, life time,
capacity factors and other operational characteristics of the plants
considered.
The values calculated by Vant-Hull (1992) are the lowest of the reviewed
studies probably due to the low emission factors associated to some materials and due to the fact that emissions associated to the operation and
maintenance activities are not included as discussed by Lenzen (1999). The
values calculated by Kreith et al. (1990) only include CO2 and do not include
© Woodhead Publishing Limited, 2012
Socio-economic and environmental assessment
125
Table 5.2 Greenhouse gas (GHG) emissions of CSP plants
GHG emissions (g CO2 equiv/kWh)
Solar only operation
Kreith et al. 1990 (Solar Two type) (I-O LCA)
Vant-Hull 1992 (Solar Two type) (Process LCA)
Norton and Lawson 1996 (Process LCA)
Röder 1997 (Phoebus type CR and SEGS type PT)
(Process LCA)
Weinrebe et al. 1998 (Phoebus type CR and SEGS
type PT) (Process LCA)
Lenzen 1999 (Solar Two type CR and ANU type PT)
(Process LCA)
Lenzen 1999 Including operation and maintenance
(Solar Two type CR and ANU type PT) (I-O LCA)
Viebahn 2003 (SEGS type PT) (Process LCA)
Viehban et al. 2008 (Solar Tres type for CR and
ANDASOL type for PT) (Process LCA)
Pehn 2006 (Process LCA)
Burkhardt et al. 2010 (Hybrid EIO LCA)
43
11
21–48
39–46
30–80
26
23–25
17
33–39
16
60
90
22
12
33
345
234
300
203
185
186
188
128
160
779
692
608
436
391
109
11
90
Solar thermal,
high
Wind, offshore
57
Solar
thermal, low
8
Solar PV,
open space
10
Solar PV, roof
11
Wind,onshore
NG GT
NGCC
Lignite IGCC
Lignite
condensing
HC IGCC
HC
condensing
LO gas
turbine
HO
condensing
14
Hydopower
ROR
209
Nuclear
GHG emissions
(g CO2 equiv/kWh)
818
726
Parabolic
trough
14
26 (24–39)
Hybrid operation
Weinrebe et al. 1998 (Process LCA)
Lenzen 1999 (Natural gas back-up capacity factor
50%) (I-O LCA)
Lechón et al. 2008 (natural gas back-up capacity
factor 71% CR and 44% PT) (Process LCA)
de la Rúa 2009 (Solar Tres type) (Process LCA)
de la Rúa 2009 (Solar Tres type) (I-O LCA)
Viebahn et al. 2008 (Solar Tres type for CR and
ANDASOL type for PT) (Process LCA)
900
800
700
600
500
400
300
200
100
0
Central
receiver
5.1 Greenhouse gas emissions of different electricity generating
technologies (source: CASES project www.feem-project.net/cases).
HO: heavy oil; LO: light oil; HC: hard coal; IGCC: integrated gasification
combined cycle; NGCC: natural gas combined cycle; NG: natural gas;
GT: gas turbine; ROR: run of river; PV: photovoltaic.
© Woodhead Publishing Limited, 2012
126
Concentrating solar power technology
the effect of other greenhouse gases, nor the emissions associated to the
operation and maintenance (O&M) activities. Values reported by Weinrebe
et al. (1998) are also quite low probably due to differences in GHG emission
intensities of materials.
The values reported by Lenzen (1999) are the highest in the range of
reported values. One of the reasons behind this, also acknowledged by the
author, could be the methodological approach followed by this author. In
fact, Lenzen’s values are calculated using an input–output methodology
which is able to capture the indirect requirements neglected by the standard
process LCA (for an introduction to I–O analysis applied to LCA, see Finnveden et al., 2009 and de la Rúa, 2009). These indirect requirements are
especially relevant when dealing with O&M activities. A standard LCA
process of a solar plant only cannot capture the emissions associated to these
activities since these emissions are indirect emissions produced in the relevant sectors of the economy such as technical services, mechanical repairs,
business services, marketing and business management, insurance, etc.
Emissions are higher in hybrid operation for the obvious reason of
fossil fuel consumption. These emissions increase with the degree of
hybridization.
The work of Lenzen (1999) showed how the GHG emissions depend on
whether the plant’s capacity factor is increased using a fossil fuel back-up
(650 g CO2 equiv/kWh at 60% capacity factor) or heat storage system (60 g/
kWh for the same 60% capacity factor). Lenzen also demonstrated that
additional capacity factor in the form of heat storage and oversized solar
field can be installed at lower marginal GHG cost than the base solar capacity itself, and that there are clear economies of scale in the GHG emissions
of both parabolic troughs and central receiver plants.
In general, the reviewed studies show that the main constituents of the
sum of the GHG emissions in solar only operation are the steel used in the
solar field, the concrete used in the solar field and in the tower and the salts
used in the storage systems.
Electricity and other fuels are consumed in the manufacturing of the
materials used to build the CSP plants. The impacts imported into the CSP
systems through these energy vectors are an example of what Pehnt (2006)
called ‘imported impacts’ which are impacts brought into the system due to
the ‘background system’ and which are not inherent to the renewable technology. These impacts can change with time as the mix of technologies used
to produce those energy vectors changes.
The relative contribution to GHG emissions of the different aspects of
the operational stage of one of the CSP plants studied by Lechón et al.
(2008) is shown in Fig. 5.2. It is worth noting the relevant contribution to
the global warming emissions of the electricity consumption of the plants
in this study. Electricity consumption in the plants is taken from the grid
© Woodhead Publishing Limited, 2012
Socio-economic and environmental assessment
127
Solar trough
plant operation
161
Natural gas
procurement
16.6
Spanish electricity
mix 2004
48.4
Electricity coal
power plant
32.8
Natural gas
burning
95.4
Electricity gas
power plant
14.72
5.2 Relative contribution to GHG emissions (g/kW h) of the different
aspects of the operational stage of a parabolic trough CSP plant
(source: Lechón et al., 2008).
instead of from the self-produced electricity. The contribution of electricity
consumption to the GHG emissions is mainly due to the fact that an important part of the electricity generation in Spain is produced in coal power
plants with very high associated greenhouse gas emissions.
Recent work by Burkhardt et al. (2010) shows the effect of different
thermal energy storage systems, cooling systems and different origin of the
salts in LCA results. According to these authors, the use of thermocline
thermal energy storage instead of a two-tank configuration allows the
reduction of 2 g CO2 equiv/kWh due to the significantly lower material
requirements. The use of a dry cooling system instead of a wet cooling
system increases GHG emissions by 2 g CO2 equiv/kWh due to the reduction in the efficiency of the steam cycle. The use of synthetic salts instead
of mined salts, considered in their base assumption, increased GHG emissions by 13 g CO2 equiv/kWh. Finally, according to their results, if both
synthetic salts and a thermocline configuration are used, the negative effects
of synthetic salts are compensated by the reduced salt requirements of the
thermocline system and the GHG emissions are increased by only 2 g CO2
equiv/kWh.
Other impacts usually assessed in LCA are acidification and eutrophization. The term acidification refers to the processes that increase the acidity
of water and soil systems through the deposition of negatively charged
ions that are then removed by leaching or biochemical processes leaving
excess H+ concentration in the system. Acidification of soils causes losses
in forest and plant health and also ecotoxicological impacts due to the
mobilization of aluminium. Acidification of water leads to loss of aquatic
© Woodhead Publishing Limited, 2012
128
Concentrating solar power technology
Table 5.3 Acidification impacts of CSP plants
Acidification (mg SO2 equiv/kWh)
Central tower
Solar only operation
Viehban 2004 (SEGS type PT)
Pehnt 2006
Hybrid operation
Lechón et al. 2008
Weinrebe et al. 1998
Parabolic trough
69.28
98
612
370
590
510
Table 5.4 Eutrophization impacts of CSP plants
Eutrophization (mg PO4 equiv/kWh)
Central tower
Solar only operation
Viehban 2004 (SEGS type PT)
Pehnt 2006
Hybrid operation
Lechón et al. 2008
Weinrebe et al. 1998
Parabolic trough
5.69
10
49.6
40
49.7
56
life (Udo de Haes et al., 2002). Eutrophization refers to the nutrient enrichment of aquatic or terrestrial environments. Aquatic eutrophization leads
to a shift of the biological structure of the aquatic environment with adverse
effects on the fauna and flora through a complex chain of ecological effects.
Terrestrial eutrophization refers to the adverse effects of excess nutrients
on plant functioning and on species composition in natural ecosystems.
With regard to acidification and eutrophization impacts, CSP systems also
show clear benefits. If electricity produced by CSP plants offsets electricity
produced by, for example, the Spanish electricity mix with an acidification
potential of around 5,000 mg SO2equiv/kWh and an eutrophization potential of around 250 mg PO4/kWh (Ecoinvent, 2007), significant impacts are
avoided.
Acidification and eutrophization impacts reported in the literature are
shown in Tables 5.3 and 5.4. In the case of CSP plants operating in solar
only mode, acidification values reported by Viebahn (2004) are 69.28 mg
SO2 equiv/kWh for a parabolic trough plant and Pehnt (2006) reported
values of 98 mg SO2 equiv/kWh. In the case of hybrid operation, the values
reported in the literature are considerably higher: 590–612 SO2 equiv/kWh
reported by Lechón et al. (2008) and 370–510 reported by Weinrebe et al.
© Woodhead Publishing Limited, 2012
Socio-economic and environmental assessment
129
Table 5.5 GHG emission factors and acidification and eutrophization potentials
of different materials
Aluminium (production mix)
Cast iron
Chromium steel
Reinforcing steel
Concrete
Copper
Glass coated
Glass uncoated
Glass tube
KNO3 as N
Ca(NO3)2 as N
GHG emission
factor (kg CO2
equiv/kg)
Acidification
potential
(g SO2 eq/kg)
Eutrophization
potential
(g NOx eq/kg)
2.9
3.1
5.3
1.5
262.6
2.2
0.7
0.6
2.5
16.1
4.0
14.5
10.4
27.1
5.2
437.0
14.9
9.4
8.5
10.9
42.4
12.7
6.8
8.9
15.7
3.8
603.7
5.7
6.2
5.5
14.7
81.4
7.1
Source: Ecoinvent (2007).
(1998). However, it is important to acknowledge that most of the impacts
are produced in the operation of the power plant due to the consumption
of natural gas and external electricity. In the case of eutrophization the
values found in the literature range from 6 to 10 mg PO4/kWh in the CSP
plants operation in solar only mode and around 50 PO4/kWh when operating in hybrid mode.
The impacts associated with the solar field are also of great importance
in some impact categories such as human toxicity and freshwater aquatic
eco-toxicity. When tracing back the origin of these impacts, it was found
that they were due to the use of steel in the metallic structure of the collectors (Lechón et al., 2008).
GHG emission factors and acidification and eutrophization potentials of
different materials used in the construction of a CSP plant are shown in
Table 5.5. As shown in Table 5.5, aluminium has lower associated environmental impacts than chromium steel, which is the material usually employed
in the construction of the metallic structure of the collectors of parabolic
trough CSP plants. The use of aluminium instead of steel in the frames as
in the case of the Acciona Solar Power SGX2 (Fernández-García et al.,
2010) space frame could reduce these impacts accordingly.
Regarding freshwater consumption of CSP plants, the work of Burkhardt
et al. (2010) provides very interesting results of a dry cooling system design.
According to these authors, the CSP plant using a wet cooling system would
consume 4.7 l of fresh water per kWh of electricity produced, while a dry
cooling system would achieve a 77% reduction in this water consumption.
© Woodhead Publishing Limited, 2012
130
Concentrating solar power technology
5.2.2
Environmental externalities assessment
of CSP systems
As explained in the introductory section of this chapter, all power generation technologies are accompanied by externalities, costs imposed on individuals or the community that are not paid for by the producer or consumer
of electricity.
The most important project on determining the externalities of energy is
the European ExternE project (www.externe.info). It was launched in 1991
by the European Commission and the US Department of Energy, and the
European Commission has been supporting the research until now through
several projects. The last one of these projects is the NEEDS Project (New
Energy Externalities Development for Sustainability, www.needs-project.
org/).
The ExternE methodology is widely accepted by the scientific community
and is considered as the world reference in the field. The quantification of
the external costs is based on the ‘impact pathway’ methodology, which was
developed in the series of ExternE projects, and has been further improved
in the NEEDS projects and other related projects like the EU CASES
project (www.feem-project.net/cases). The impact pathway methodology
aims at modelling the causal relationships from the emission of a pollutant
to the impacts produced on various receptors through the transport and
chemical conversion of this pollutant in the atmosphere. The main receptors
of the impacts are human health, crops, ecosystems and materials. Welfare
losses produced by these impacts are assessed using economic valuation
methods. Impact categories, pollutants and effects considered in the ExternE
methodology are summarized in Table 5.6.
Global warming impacts assessment is subject to a high degree of uncertainty. Within NEEDS, the model FUND 3.0 was used to estimate the
marginal external costs of GHG emissions (Anthoff, 2007). Results differ
greatly depending on the assumptions regarding some very influential
parameters such as discounting and equity weighting. Two sets of external
costs factors were used in NEEDS, trying to reflect these uncertainties
(Preiss and Friedrich, 2009).
The results of the external costs assessment of CSP plants from the
NEEDS project are shown in Fig. 5.3. External costs other than global
warming costs decrease with time as the technology matures and the inventories of pollutant emissions decrease. However, marginal external costs for
GHG emissions increase with time, and therefore the total external costs
of CSP systems increase. External costs calculated for CSP technology are
quite low compared with other competing electricity generation technologies as shown in Fig. 5.4. Fossil fuel technologies have external costs above
1.4 eurocent/kWh. These costs are dominated by global warming impacts
in the case of coal, lignite and natural gas and by health effects in the case
© Woodhead Publishing Limited, 2012
Socio-economic and environmental assessment
131
Table 5.6 Impacts, pollutants and effects covered by the ExternE methodology
Impact category
Pollutant
Effects
Human health
– mortality
PM10, SO2, NOx, O3
As, Cd, Cr, Ni
Accident risk
Human health
– morbidity
PM10, O3, SO2
PM10, O3
PM10, CO
PM10
Reduction in life expectancy
Cancer
Fatality risk from traffic and
workplace accidents
Respiratory hospital admissions
Restricted activity days
Congestive heart failure
Cerebro-vascular hospital
admissions
Cases of chronic bronchitis
Cases of chronic cough in
children
Cough in asthmatics
Lower respiratory symptoms
Neurotoxicity
Asthma attacks
Symptom days
Cancer risk (non-fatal)
Pb
O3
Benzene
Benzo-[a]-pyrene
1,3-butadiene
Diesel particles
Noise
Accident risk
Building materials
SO2
Acid deposition
Combustion particles
Crops
NOx, SO2
O3
Global warming
Ecosystems
Acid deposition
CO2, CH4, N2O, N, S
Acid deposition
Nitrogen deposition
Myocardial infarction
Angina pectoris
Hypertension
Sleep disturbance
Risk of injuries from traffic and
workplace accidents
Ageing of galvanized steel,
limestone, mortar, sandstone,
paint, rendering, and zinc for
utilitarian soiling of buildings
Yield change for wheat, barley,
rye, oats, potato, sugar beet
Yield change for wheat, barley,
rye, oats, potato, rice,
tobacco, sunflower seed
Increased need for liming
World-wide effects on mortality,
morbidity, coastal impacts,
agriculture, energy demand,
and economic impacts due to
temperature change and sea
level rise
Acidity and eutrophication
(avoidance costs for reducing
areas where critical loads are
exceeded)
Source: EC (2005).
© Woodhead Publishing Limited, 2012
External costs (cEuro/kWh)
132
Concentrating solar power technology
0.50
0.42
0.40
0.30
0.20
0.24
0.20
0.20
0.20
Today
2025
0.22
0.10
0.00
Today
2025
2050
Scenario I
Health impacts
Material damage
Global warming Scenario II
2050
Scenario II
Biodiversity
Land use
Crop yield losses
Global warming Scenario I
5.3 External cost evolution of CSP systems (source: authors’
elaboration using data from NEEDS project).
of heavy and light oil. Among fossil technologies, the ones with higher
efficiencies have correspondingly less external costs per kWh. Nuclear
energy external costs are very low in these calculations, but do not take into
account the effects of a possible nuclear accident or the effects on the future
environment and society of the possible accidental release of the nuclear
waste that has been disposed of (Lecointe et al., 2007). Solar PV technologies have sensibly higher external costs than CSP and these costs are dominated by the health effects arising from the emissions originated by the
energy requirements of the upstream processes related to the production
of silicon and PV wafers (Frankl et al., 2006). Improvements in energy
consumption and also the better efficiencies that are foreseen for this technology would reduce the external costs accordingly.
5.3
Socio-economic impacts of concentrating solar
power (CSP) systems
The benefits associated with solar thermal electricity deployment are
various in nature and should be taken into consideration in order to design
support policies aimed at compensating its higher electricity production
costs compared to fossil fuel alternative technologies. As described in detail
in the previous section, among other environmental impacts associated with
the gradual substitution of fossil fuel technologies by CSP technologies, CO2
emissions as well as energy consumption reductions are some of the most
notable benefits (Lechón et al., 2008). However, other socio-economic
impacts should equally be taken into consideration.
© Woodhead Publishing Limited, 2012
© Woodhead Publishing Limited, 2012
0.00
0.50
1.00
1.50
2.00
2.50
3.00
3.50
HC IGCC
HC condensing
LO gas
turbine
NGCC
Lignite IGCC
Lignite
condensing
Environmental impacts
Hydropower,
dam
Hydropower,
ROR
NG GT
Radiological impacts
Wind, offshore
Wind, onshore
Hydropower,
PS
Global warming impacts
Solar PV,
open space
Solar PV, roof
HO
condensing
Nuclear
5.4 External costs of different electricity generating technologies (source: authors’ elaboration using CASES project data,
www.feem-project.net/cases).
Human health impacts
Solar thermal
External costs (cEuro/kWh)
134
Concentrating solar power technology
As previously mentioned, compared to fossil and other renewable energy
technologies, one of the most relevant of CSP’s associated socio-economic
impacts is its capacity to stimulate the economy and create new jobs at the
local level. One of the main reasons for this is that CSP’s ‘high-tech’ component requirements are low and its main components include steel, concrete, mirrors and labour. Such local effects may be realized through an
increase in the demand for goods and services as well as through the creation of new jobs. These impacts may take the form of:
•
•
•
direct effects – accrued due to the increase in the demand for those
industries that directly provide goods and services required to construct,
operate, maintain and dismantle a plant,
indirect effects – originated due to the effect that such new investment
has on new flows of purchases and/or sales among other productive
sectors in the economy, and
induced effects – related to the expansion of private expenditure in
goods and services (food transportation, health, services, etc.) from the
workers employed – in a direct or indirect way – by the project.
In that sense, Kulstic et al. (2007) highlight the fact that many assessments
currently underestimate the total socio-economic effects since they only
take into account the direct effects and disregard induced and most important indirect effects that take place during the construction, operation,
maintenance and dismantling of any power plant. In order to fully account
for the impacts on the demand for goods and services as well as on employment, one of the soundest analytical tools is the input–output methodology,
which will be presented below. After describing the methodology as well as
highlighting its main advantages and limitations, the following section will
present an example of its application used to estimate the socio-economic
impacts associated with the solar thermal energy deployment in Spain.
5.3.1
Input–output methodology
The input–output (I-O) methodology, which was first developed by Wassily
Leontief in the late 1930s, has been widely used to trace out a portrait of
the whole national economic structure (Leontief, 1966, 1986). The input–
output symmetric table is an economic analysis tool that reflects the value
of the different goods and services that are exchanged in an economy. The
structure of the I-O table is such that, along the different rows and columns
of the matrix, one can find the different sectors within the economy set in
a symmetrical way. The different elements displayed along each row describe
the different uses of each sector’s production. In a similar manner, for each
sector in the economy, the elements along the columns of the symmetric
input–output table account for the resources that have been consumed from
© Woodhead Publishing Limited, 2012
Socio-economic and environmental assessment
135
Table 5.7 Input–output table structure
Resources (columns)
Uses (rows)
1.
2.
3.
4.
1.
2.
3.
4.
5.
6.
Intermediate resources
Value added
Effective production (1 + 2)
Imports
5. Total resources ( 3 + 4)
Intermediate products
Final consumption
Gross capital formation
Exports
Final uses (2 + 3 + 4)
Total uses (1 + 5)
Table 5.8 I-O symmetric table scheme
Sectors’ consumption
Sectors’ production
1
2
3
n
Intermediate
consumption I
Added value V
Total production X
1
X11
X21
X31
Xn1
I1
2
X12
X22
X32
Xn2
I2
3
X13
X23
X33
Xn3
I3
n
X1n
X2n
X3n
Xnn
In
V1
X1
V2
X2
V3
X3
Vn
Xn
Final
demand Y
Total
production X
Y1
Y2
Y3
Yn
X1
X2
X3
Xn
GDP
Source: Hendrickson et al. (2006).
other sectors in order to obtain a certain production in each sector. In a
schematic and simplistic way, an I-O table can be depicted as in Table 5.7.
Based on the input–output symmetric table shown in Table 5.8, the matrix
of coefficients summarizes the interdependencies between production
sectors (Ten Raa, 2005) and is used to analyse the economic activity and
employment impacts induced by an increase in the demand of any particular economic sector.
Increase in the demand for goods and services
According to the I-O methodology, the relationship between the expenditure generated by a certain project and its impact on the demand for goods
and services is depicted by the following relation:
ΔQy = ( I − A) ΔDy
−1
[5.1]
where: ΔQy = increase in the total demand for goods and services (direct
and indirect), I = matrix unit, A = technical coefficients matrix, and ΔDy =
© Woodhead Publishing Limited, 2012
136
Concentrating solar power technology
increase in direct demand for goods and services generated by the development of a certain project y.
The different elements included in matrix A (n × n) are named technical
coefficients (aij) which reflect the percentage of goods or services from
sector ‘i’ that are required to produce one good or service unit from sector
‘j’. Put another way, the technical coefficients indicate the amount that
sector ‘j’ requires from sector ‘i’ in order to produce one unit of product or
service j (both quantities should be expressed in their monetary value at
constant prices).
A = {aij }, being aij =
xij
xj
[5.2]
where xij = goods or services that sector j requires from sector i (in monetary
terms), and xj = total production from sector j.
In most countries, an official institution regularly publishes the technical
coefficient matrix A (as well as the symmetric input–output table, upon
which the technical coefficient matrix is built). In the case of Spain, every
five years the National Statistics Institute (INE) publishes the National
Input–Output tables based on the National Accounts records. Once the
so-called Leontief inverse matrix (I-A)−1 has been constructed, it is then
possible to estimate the impact derived from a certain project by multiplying (I-A)−1 by the investment as well as operation and maintenance costs
vector ΔD associated to the project. The result from this operation is a
column vector ΔQ(n ∗ 1) the sum of whose elements is the total impact of
the investment, which includes both direct and indirect impacts.
Employment creation
Besides increasing the demand for certain goods and services, the development of any project generates impacts on the employment in a direct and
indirect way. In order to estimate such effect, a column vector Ls must be
constructed based on the number of employed people across each sector
in the economy (expressed as number of employed people for every million
euros produced in each sector). Secondly, Ls must be multiplied by ΔQ
(which represents the previously obtained vector that accounts for the total
economic impact). The result from this multiplication is the total number
of employments that have been created in each sector due to this project.
Each element of the resulting vector shows the total number of new jobs
in each sector created both in a direct or indirect way.
Ls ΔQy = Ly
[5.3]
where Ls = vector of employees per sector, and Ly = direct and indirect
impact of employment due to the project. Based on this estimation, the
© Woodhead Publishing Limited, 2012
Socio-economic and environmental assessment
137
number of direct and indirect employments could be estimated using the
following expressions:
Direct employment = Ls ΔDy;
indirect employment = Ls (ΔQy − ΔDy )
Finally, and based on the previous results, it is then possible to compute the
multiplying effect of a certain project. A multiplier is a number that indicates by how much a certain economy is going to grow due to a certain
project development (taking into account both direct and indirect effects).
The general formula to compute the multiplying effect (M) is:
Multiplier (M ) =
Total effects ΔQ
=
Direct effects ΔD
[5.4]
Compared to other alternative analytical methods, the most relevant advantages of the input–output methodology are its simplicity, intuitive understanding, basic software requirements as well as its acceptability within the
scientific community. However, among its limitations, it is worth mentioning
that its constant technical coefficients do not always take into account
technological improvements, import substitution, change in consumption
patterns or relative price variations that take place from one year to another
(Holland and Cooke, 1992). Moreover, homogeneity among sectors as well
as lack of production capacity limitations is assumed.
Despite the above mentioned limitations, the input–output methodology
has been widely applied to study the socio-economic impacts of various
energy projects (Tegen, 2006; Lantz and Tegen, 2009; Linares et al., 1996;
Caldés et al., 2009; de la Rúa, 2009; Lanier et al., 1998). In that sense, in a
recent study by the National Renewable Energy Laboratory (Lantz and
Tegen, 2009), the input–output methodology is presented as one of the most
consolidated methods recognized by the scientific community.
5.3.2
Application of an input–output analysis: estimation
of the socio-economic impacts of CSP energy
deployment in Spain
Over the last few years, Spain’s solar thermal electricity deployment has
been remarkable, mainly due to its regulatory environment as well as
favourable climatic conditions. This favourable context has brought
forth an upsurge in solar projects – mainly using either a central receiver
or parabolic trough technologies – and it is expected that, in the near future,
the potential CSP capacity in Spain will exceed 500 MW; the Spanish
Renewable Energy Plan (PER) goal for solar thermal installed capacity by
2010.
© Woodhead Publishing Limited, 2012
138
Concentrating solar power technology
In this context, the goal of the work by Caldés et al. (2009) that will be
presented here was to estimate the socio-economic impacts of increasing the
installed solar thermal energy power capacity in Spain by using an input–
output analysis. For more detailed information, see Caldés et al. (2009).
In order to estimate such effects, two scenarios were considered:
1. The first scenario considered the individual impacts derived from the
construction and operation of two solar thermal power plants with the
following specifications:
• A 50 MW power plant consisting of 624 parabolic trough collectors.
This plant uses synthetic oil as the heat transfer fluid and molten salts
to provide seven hours’ storage at peak output. Following the current
regulatory framework, 15% of total output is generated by natural gas.
• A 17 MW central solar tower power plant consisting of 2,750 heliostats. This plant uses molten salts both as a heat transfer fluid and
storage system. This power plant occupies 150 ha and, as in the previous case, the power plant generates 15% of electricity from natural
gas.
2. The second scenario replicates the PER installed capacity goal for solar
thermal power by 2010 which would lead to 500 MW installed capacity.
According to this hypothetical scenario, it was assumed that 80% of such
capacity would be met by parabolic trough plants, while 20% would be
met by solar tower power plants.
Solar thermal plant costs
Based on actual cost data of CSP projects currently in operation in Spain,
this section presents a summary of the main data and assumptions used to
construct the plant as well as O&M cost vectors associated to each of the
power plants analysed. It must be noted that, due to the lack of precise data
on employment and salary figures, induced effects were not estimated. In
the same way, due to the lack of data, the end-of-life dismantling phase of
the project was not taken into account.
Parabolic trough power plant (50 MW)
Of the total investment costs (265,837 kc), the solar field accounts for 46%
of the total investment cost, power block 21%, storage 13%, construction
10% and the remaining 10% is accounted for by engineering costs and
contingencies. With respect to the annual operation and maintenance costs
(12,300 kc), an operational lifetime of 25 years with an annual discount rate
of 8% was considered. Within total operation and maintenance costs
(240,380 kc over the life of the plant), it was assumed that the payment of
employees’ wages account for 80% of such costs (representing 1,033.6 kc/
© Woodhead Publishing Limited, 2012
Socio-economic and environmental assessment
139
year), while the rest is accounted for by administration services, insurance,
etc. Consequently, and given that the average salary of a Spanish employee
working in the electricity generation and distribution sector amounts to
46.3 kc/year (INE, 2006), the estimated number of people annually working
in the operation of this plant would amount to 22 people. Finally, expenses
associated with natural gas and electricity consumption were also accounted
for and the estimation of the financing expenses has been computed assuming a 12-year repayment loan with a 7% interest rate.
Solar tower power plant (17 MW)
Out of the total investment cost of the plant (147,016 kc), it was assumed
that the solar field accounts for 43% of the investment, power block 20%,
tower and receptor 16%, storage system 6%, construction 7% and the
remaining 8% is accounted for by engineering and contingencies costs. With
regard to its annual operation and maintenance costs (7,154 kc/year), it was
assumed that the operational period lasted 25 years with a 2% annual discount rate. As in the previous plant, within the total fixed operation and
maintenance costs (139,811 kc over the life of the plant), it was assumed
that 80% of such costs accrued to employees’ wages and the rest was
accounted for by administrative services, insurances, etc. As in the previous
case, the estimated number of people annually employed in the plant
amounted to 22 people. Both gas and electricity costs required to operate
the solar plant were taken into account and investment costs were financed
with a loan to be repaid over 12 years with an annual interest rate of 7%.
National economic data
At the time when this study was conducted, the most up-to-date official
Spanish input–output table was used. This 2000 I-O table was published by
the National Statistics Institute in 2007 and reflected all transactions that
had taken place across economic sectors in the form of increased demands
as well as intermediate and final production across 73 national economic
sectors. Based on this original I-O table, a reduced table consisting of the
22 most relevant economic sectors for this analysis was constructed. Finally,
total costs associated with the studied plants had to be broken down later
and associated with the different sectors in the reduced I-O table.
Results: parabolic trough plant
As shown in Table 5.9, the total effect associated with this plant amounts
to 930 Mc (equivalent to 18.6 Mc/MW), of which total indirect effect generated during the construction and operational phase accounts for 445 Mc.
© Woodhead Publishing Limited, 2012
140
Concentrating solar power technology
Table 5.9 Total effect on the demand for goods and services as well as
employment per MW of installed capacity for 50 MW parabolic trough power
plants (PTP)
Effect on demand for goods and
services
Demand ratio
(Mc/MW)
Direct demand (Const + Operation)
Indirect demand
(Const + Operation)
Total increase in demand
(Const + Operation)
Total effect (Mc)
9.7
8.9
485
445
18.6
930
Effect on employment creation
Direct employment
(Const + Operation)
Indirect employment
(Const + Operation)
Total employment
Employment ratio
(person years/MW)
Total employment
(person years)
111
5,553.5
81
4,030.2
192
9,583.7
The associated multiplier effect is 1.92 which means that for every euro
invested during the construction and operation phase of the plant, an aggregate demand of 1.92 euros is generated.
With respect to its effect on employment creation, the above mentioned
increased demand for goods and services would generate 9,583.7 additional
person years of employment (of which 5,553.5 and 4,030.2 are directly and
indirectly created, respectively). This figure implies that for every 19.8 thousand c directly invested, one person year of employment is created.
Results: solar tower plant
With respect to the socio-economic impacts associated to the solar tower
plant (Table 5.10), the total effect on the demand for goods and services
amounts to 521.9 Mc which is equivalent to 30.7 Mc/MW. Of this total
effect, the indirect effect generated during the construction and operation
phase accounts for 256 Mc and the total multiplier effect is 1.96.
With respect to the employment effect, during the construction and operation of the plant, 5,491 person years of employment would be created
© Woodhead Publishing Limited, 2012
Socio-economic and environmental assessment
141
Table 5.10 Total effect on the demand for goods and services as well as
employment per MW of installed capacity for 17 MW solar tower power plant
(TP)
Effect on demand for goods and
services
Direct demand (Const + Operation)
Indirect demand
(Const + Operation)
Total increase in demand
(Const + Operation)
Demand ratio
(Mc/MW)
Total effect (Mc)
15.7
15
266
256
30.7
522
Effect on employment creation
Direct employment
(Constr + Operation)
Indirect employment
(Constr + Operation)
Total employment
Employment ratio
(person years/
MW)
Total employment
(person years)
189
3,213
133
2,278
322
5,491
(3,213 directly and 2,278 indirectly) implying that one new person year of
employment is created for every 20.6 thousand euros directly invested.
Compliance with the PER objectives
Based on these individual plant results, the socio-economic impacts associated to the compliance with the Spanish Renewable Plan 2005–2010 (PER)
were estimated. It was assumed that in order to meet the 2010 PER solar
thermal installed capacity goal (500 MW installed capacity), 400 MW of
parabolic trough plants (80% of the total power) would be installed, while
the rest (20%) would be met with 100 MW of solar thermal tower plants.
Furthermore, it was assumed that during the period under consideration,
operation and investment costs would remain constant, an assumption supported by the literature (DLR, 2005; DLR et al., 2005).
Results show that the total increase in the demand for goods and services
generated as a result of compliance with the PER solar thermal goal would
amount to 10,538 Mc (equivalent to an average of 21 Mc/MW) (Table 5.11).
© Woodhead Publishing Limited, 2012
142
Concentrating solar power technology
Table 5.11 Total effect on the demand for goods and services as well
as employment per MW of installed capacity for 500 MW PER 2005–
2010 solar thermal goal
Effect on demand for goods and
services
Total increase in demand
(Const + Operation)
Demand ratio
(Mc/MW)
Total effect (Mc)
21
10,538
Effect on employment creation
Total employment
(Const + Operation)
Employment ratio
(person years/
MW)
Total employment
(person years)
218
108,992
Source: Ragwitz et al. (2009).
Total employment would amount to 108,992 person years of employment
(63,485 direct jobs and 45,508 indirect jobs). Results show that the total
effect on the demand for goods and services would amount to 10,538 MW
which is equivalent to an average of 21.1 Mc for every MW installed.
With respect to the effect on the job creation, total direct employment
generated would be 63,485 person years while the indirect employment generated would reach 45,508.
Although results should be interpreted with caution due to the methodological limitations of the input–output methodology, it can be concluded
that the socio-economic benefits derived from the accomplishment of the
PER’s solar thermal installed capacity in Spain would be remarkable, both
in terms of increased demand for goods and services as well as employment
creation.
While few studies have specifically analysed the impact of CSP technologies on employment (Caldés et al., 2009; de la Rúa, 2009), there exist other
reports that simultaneously estimate the employment impact derived from
the deployment of various renewable energy technologies at the national
level (GFME, 2006; Hillebrand et al., 2006; APPA/Deloitte Consulting,
2009). At the European level, it is worth mentioning the work by Whitely
et al. (2004) and Ragwitz et al. (2009). The latter is one of the most comprehensive studies in this area since it estimates the gross and net effect
© Woodhead Publishing Limited, 2012
Socio-economic and environmental assessment
143
Table 5.12 EU economic and employment impact of RES deployment in 2005
Gross value added
Employment
Employment in SME
Employment in agriculture/
forestry
Relative impacts:
Gross value added compared
to total GDP
Employment compared to
total employment
bin. c2,000
min. employees
min. employees
min. employees
Direct
impact
Indirect
impact
Total
impact
31.9
0.8
0.5
26.4
0.6
0.4
58.3
1.4
0.9
0.2
0.32%
0.58%
0.36%
0.64%
Source: Ragwitz et al. (2009).
W
in
ot
d
ov
la
o
lta
rt
he
ic
s
rm
al
he
G
at
eo
th
er
m
H
al
ea
tp
um
ps
So
Ph
ls
La
rg
e
hy
dr
Sm
o
al
lh
yd
ro
ue
te
of
Bi
ow
as
gr
id
n-
no
Bi
gr
id
s
om
as
s
RES fuel use
RES operation
RES investment
Bi
Bi
om
as
Bi
og
as
1,000 employed persons
500
450
400
350
300
250
200
150
100
50
0
5.5 EU economic and employment impact of RES deployment in 2005
(source: Ragwitz et al., 2009).
that the different renewable energy (RE) technologies have on both
employment generation as well as on the economy at the EU-27 level in
2005 and under two different future RE deployment scenarios (see Table
5.12 and Fig. 5.5).
5.4
Future trends
The role of solar technologies in a future energy supply system remains
uncertain but, according to various experts, due to the expected future
© Woodhead Publishing Limited, 2012
144
Concentrating solar power technology
technology innovation and developments, it indeed looks promising (Nuño,
2008). With regard to CSP, and according to various experts, over the next
50 years, constructions, functional surfaces, mirrors, heat transfer media and
systems design will be greatly optimized in regard to costs, efficiency and
environmental impact. It is expected that part of the cost reduction potential and efficiency increase should be realized by R&D, scale effects and
volume effects (Viebahn et al., 2008).
5.4.1
Projections of environmental impacts
The evolution of the GHG and other emissions over time is evaluated in
the work performed in the NEEDS project. When assessing the emissions
of future configurations of CSP power plants, a clear reduction is observed
showing an ‘environmental learning’ of the technology (see Fig. 5.6). The
three development scenarios, very optimistic, optimistic-realistic and pessimistic, refer to different assumptions on the anticipated penetration of the
technology into the energy market, reaching an installed capacity in 2050
of 1,000 GW, 405 GW and 120 GW, respectively. In these scenarios the
prevailing CSP technology differs. Greenhouse gas emissions show a large
reduction throughout the scenario development. The reason is the reduction of salt used in the different storage systems. Concrete storage or PCM
storage-based power plants perform better than the current molten saltbased ones. The results show a continuous optimization from the ‘pessimistic’ to the ‘very optimistic’ scenario as well as over time.
5.4.2
Projections of socio-economic impacts
As explained previously, some of the most relevant CSP socio-economic
impacts include their potential to stimulate the local or national economy
by increasing the demand for goods and services as well as to generate new
jobs. The magnitude of such effects will greatly depend on the future evolution and characterization of CSP’s costs as well as on the location and
labour intensity of such investments.
With regard to the 2050 CSP cost development target, several studies and
models expect to reach 4–5 ct/KWh for base load, 5–8 ct/KWh for mid-load
and more than 10 ct/KWh for pure peak load. Various factors affect future
evolution of the resulting levelized cost of energy (LCOE), which is the most
important determinant of the assumed development of the specific investment costs for new plants. This value comprises the assumptions about cost
reduction due to technical learning, scale-up and volume effects (Viebahn et
al., 2008). Nevertheless, according to the European Solar Thermal Electricity
Association (ESTELA), only a moderate reduction of LCOE can be
expected due to high increase of raw materials such as steel and concrete.
© Woodhead Publishing Limited, 2012
© Woodhead Publishing Limited, 2012
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
‘Pessimistic’
‘Optimistic-realistic’
2025
Construction
Fuel
‘Very optimistic’
Operation
‘Pessimistic’
Dismantling
‘Optimistic-realistic’
2050
5.6 Evolution of the GHG emissions of CSP technologies (source: Viebahn et al., 2008).
Today
GHG emissions - 440 ppm
Parabolic tr.
salt
Tower salt
Weighted
mean
Parabolic tr.
salt
Parabolic tr.
concrete
Weighted
mean
Parabolic tr.
PCM
Fresnel tr.
PCM
Tower salt
Weighted
mean
Fresnel tr. CHP
cooling
Fresnel tr. CHP
desalting
Weighted
mean
Parabolic tr.
salt
Parabolic tr.
concrete
Weighted
mean
Parabolic tr.
PCM
Fresnel tr.
PCM
Tower salt
Weighted
mean
‘Very optimistic’
Fresnel tr. CHP
cooling
Fresnel tr. CHP
desalting
Weighted
mean
146
Concentrating solar power technology
Table 5.13 CSP investment cost projections
‘Today’
NEEDS (Viebahn et al., 2008)
Global cumulated capacity (GW)
Investment costs in c 2005/kW
2020
0.4
5,300
2025
2030
2040
63
3,720
2050
405
2,770
Energy Technology Perspectives (Blue scenario) (International Energy
Agency, 2008)
Global cumulated capacity (GW)
250
Investment costs in c 2005/kW
3,620
630
Energy Revolution (Greenpeace International/European Renewable Energy
Council, 2008)
Global cumulated capacity (GW)
Investment costs in c 2005/kW
6,000
4,170
3,530 3,480 3,440
Leitstudie (BMU, 2008)
Investment costs in c 2005/kW
3,600
3,300 3,200
3,100
Source: German Federal Environmental Agency (2009).
With regard to CSP investment cost projections, a recent study by the
German Federal Environmental Agency (2009) made a comparison of four
cost projections derived from various energy scenarios (see Table 5.13).
Despite the future LCOE, cost decline will be responsible for lower socioeconomic impacts by MW, the overall contribution to the economy and
employment will grow due to the expected installed capacity increase.
According to the four studied future energy scenarios, the investment cost
evolution is promising. However, as highlighted by the German Federal
Environmental Agency, a number of aspects should be take into account. In
the first place, thermal storage system is a key aspect which greatly affects
investment costs, not only due to the direct costs for the thermal storage itself,
but also because of the effects on the overall power plant configuration (i.e.
larger collector field). However, one must take into account that while
storage increases cost per installed kW, the capacity factor goes up and the
associated LCOE could be lower. Secondly, it is important to highlight the
fact that given the limited number of CSP commercial power plants currently
in operation, the analysed studies present learning curves which are not
based on historical data but rather apply the concept of learning curves and
an assumed learning curve (German Federal Environmental Agency, 2009).
5.4.3
Location of future CSP plants
As previously stated, one of the advantages of CSP technologies is that they
take advantage of an inexhaustible resource which is widely abundant in
© Woodhead Publishing Limited, 2012
Socio-economic and environmental assessment
147
various developed and developing countries around the world (Viebahn
et al., 2008). Given their high solar radiation, various countries are expected
to invest in new CSP capacity in the future. The most promising regions
include: Southeast of the United States, Central and South America, Africa,
Middle East, European countries along the Mediterranean, Iran, Pakistan
and the desert areas of India, as well as the former USSR, China and Australia (Aringhoff et al., 2005).
Consequently, materializing the deployment of CSP technologies in these
countries, and in particular in least developed countries, could represent a
unique opportunity to promote their development by creating new employment opportunities as well as by stimulating the national economy.
5.5
Summary and conclusions
While there exists a vast literature regarding current and future CSP technological aspects, literature that focuses on CSP’s environmental and socioeconomic assessments is relatively scarce. It is expected that in the near
future this knowledge gap will be reduced since both socio-economic and
environmental factors are going to play a key role in determining future
energy scenarios and therefore in promoting CSP deployment at the global
level. In order to improve the competitiveness of renewable energy technologies in the energy market, it is required that support policies take into
account both their environmental and socio-economic benefits.
This chapter has reviewed the existing LCA and external costs studies
on CSP technologies. Values of global warming emissions in solar only
operation reported in the literature range from 11 g/kWh to 60 g/kWh and
from 12 g/kWh to 90 g/kWh for central tower and parabolic troughs, respectively. These emissions are well below the emissions produced by conventional electricity generation sources. With regard to acidification and
eutrophization impacts, CSP systems also show clear benefits. Results of the
external costs assessment of CSP plants are around 0.2 cEuro/kWh, well
below those of other competing electricity generation technologies.
The current and future estimated socio-economic impacts of CSP technologies are not negligible. Compared to conventional fossil fuel technologies, the impact of the deployment of CSP technologies on the national and
local economy as well as on employment can be remarkable. This is particularly the case for countries such as Spain where its favourable climatic
conditions and support policies have generated an upsurge in CSP projects.
By conducting an input–output analysis, the work presented in this chapter
has attempted to first estimate the direct and indirect socio-economic
impacts associated with the construction as well as O&M phases of two
individual solar thermal power plants – a 50 MW parabolic trough plant
and a 17 MW tower plant. Based on the former results, the Spanish Plan
© Woodhead Publishing Limited, 2012
148
Concentrating solar power technology
de Energías Renovables (PER 2005–2010), the associated socio-economic
impacts (which states that by 2010 solar thermal installed capacity should
reach 500 MW) have been estimated. Results show that the total effect on
the demand for goods and services would amount to 10,538 Mc which is
equivalent to c21.1 million for every installed MW. With respect to its effect
on employment, 108,992 new jobs would be created.
5.6
References
Anthoff, D. (2007) Report on marginal external costs inventory of greenhouse gas
emissions. NEEDS Deliverable D5.2, RS1b.
APPA/Deloitte Consulting (2009) Impacto macroeconómico de las energías renovables en España.
Aringhoff, R., G. Brakmann, M. Geyer and S. Teske (2005) Concentrated Solar
Thermal Power Now. ESTELA/IEA SolarPACES/Greenpeace.
BMU (2008) Leitstudie 2008: Weiterentwicklung der ‘Ausbaustrategie Erneuerbare
Energien’ vor dem Hintergrund der aktuellen Klimaschutzziele Deutschlands und
Europas. Federal Ministry for the Environment, Nature Conservation and Nuclear
Safety, Berlin.
Burkhardt, J., G. Heath and C. Turchi (2010) Life Cycle Assessment of a Parabolic
Trough Concentrating Solar Power Plant. NREL. Available at: http://www.
lcacenter.org/LCAX/presentations-final/121.pdf
Caldés, N., M. Santamaría, M. Varela and R. Sáez (2009) Economic impact of solar
thermal electricity deployment in Spain. Energy Policy 37, 1628–1636.
Carter, N. and R. Campbell (2009) Water Issues of Concentrating Solar Power (CSP)
Electricity in the U.S. South West. Congressional Research Service.
DLR (2005) MED-CSP Concentrating Solar Power for the Mediterranean Region–
Executive Summary. Available at: www.dlr.de/tt/med-csp (accessed March 2010).
DLR et al. (2005) ECOSTAR European Concentrated Solar Thermal Road Mapping.
SES6-C6-2003-502578. Edited by R. Pitz-Paal, J. Dersch and B. Milow.
de la Rúa, C. (2009) Desarrollo de la herramienta integrada ‘Análisis de ciclo de
vida-input output’ para España y aplicación a tecnologías energéticas avanzadas,
PhD Thesis, Universidad Politécnica de Madrid.
EC (2005) ExternE, Externalities of Energy. Methodology 2005 Update. EUR
21951.
Ecoinvent (2007) Swiss Centre for Life Cycle Inventories (Ecoinvent Centre).
Ecoinvent Database. Ecoinvent Centre, Dübendorf, 2004 and 2007. Available at:
http://www. ecoinvent.org (accessed March 2010).
Fernández-García, A., E. Zarza, L. Valenzuela and M. Pérez (2010) Parabolic-trough
solar collectors and their applications. Renewable and Sustainable Energy Reviews
14, 1695–1721.
Finnveden, G., M. Z. Hauschild, T. Ekvall, J. Guinée, R. Heijungs, S. Hellweg,
A. Koehler, D. Pennington and S. Suh (2009) Recent developments in life cycle
assessment. Journal of Environmental Management 91, 1–21.
Frankl, P., E. Menichetti and M. Raugei with contributions by S. Lombardelli and
G. Prennushi (2006) NEEDS. New Energy Externalities Developments for Sustainability Deliverable 11.2 – RS 1a ‘Final report on technical data, costs and life
cycle inventories of PV applications’.
© Woodhead Publishing Limited, 2012
Socio-economic and environmental assessment
149
German Federal Environmental Agency (2009) Role and Potential of Renewable
Energy and Energy Efficiency for Global Energy Supply.
GFME (German Federal Ministry for the Environment) (2006) Nature conservation and nuclear safety. Renewable energy: employment effects. Impact of the
Expansion of Renewable Energy on the German Labour Market. Summary.
Greenpeace International/European Renewable Energy Council (2008) Energy
Revolution. A Sustainable Global Energy Outlook.
Guinée, J. B., M. Gorrée, R. Heijungs, G. Huppes, R. Kleijn, A. de Koning, L. van
Oers, A. Wegener Sleeswijk, S. Suh, H.A. Udo de Haes, H. de Bruijn, R. van Duin
and M.A.J. Huijbregts (2002) Handbook on Life cycle Assessment – Operational
guide to the ISO standards. I: LCA in perspective. IIa: Guide. IIb: Operational
annex. III: Scientific background. Kluwer Academic Publishers, Dordrecht.
Hendrickson, C. T., L. B. Lave, et al. (2006) Environmental Life-Cycle Assessment of
Goods and Services: An input-output approach. Resources for the Future,
Washington, DC.
Hillebrand, H., G. Buttermann, J.M. Behringera and M. Bleuel (2006) The expansion
of renewable energies and employment effects in Germany. Energy Policy 34,
3484–3494.
Holland D. and S. Cooke (1992) Sources of structural change in the Washington
economy: an input–output perspective. Annals of Regional Science 26, 155–170.
INE (Instituto Nacional de Estadísticas) (2006) Available at www.ine.es. (accessed
March, 2010).
International Energy Agency (1998) Benign Energy? The Environmental Implications of Renewables. IEA, Paris.
International Energy Agency (2008) Energy Technology Perspectives 2008 – Scenarios and Strategies to 2050. IEA, Paris.
ISO (2006a) ISO 14040:2006. Environmental management: Life cycle assessment –
principles and framework.
ISO (2006b) ISO 14044:2006. Environmental management: Life cycle assessment –
requirements and guidelines.
JRC IES (2010) ILCD Handbook, General Guide for Life Cycle Assessment –
Detailed Guide. Luxembourg: Publications Office of the European Union.
Kreith, F., P. Norton and D. Brown (1990) A comparison of CO2 emissions from
fossil and solar power plants in the United States. Energy 15, 1181–1198.
Kulstic, B., R. Loizou and V. Segon (2007) Impact of biodiesel production on Croatian economy. Energy Policy 35(12), 6036–6045.
Lanier, J., S. Bernow and J. Decicco (1998) Employment and other macroeconomic
benefits of an innovation-led climate strategy for the United States, Energy Policy
25(5), 425–433.
Lantz, E. and S. Tegen (2009) NREL response to the report ‘Study on the effects on
employment of public aid to Renewable Energy Sources’ from King Juan Carlos
University (Spain), White Paper, NREL/TP 6A2-46261.
Lecointe, C., D. Lecarpentier, V. Maupu, D. Le Boulch and R. Richard (2007)
NEEDS. New Energy Externalities Developments for Sustainability. D 14.2 – RS
1a ‘Final report on technical data, costs and life cycle inventories of nuclear power
plants’.
Lechón, Y., C. de la Rúa and R.M. Sáez (2008) Life cycle environmental impacts of
electricity production by solar thermal power plants in Spain. Journal of Solar
Energy Engineering 130, 021012-1–021012-7.
© Woodhead Publishing Limited, 2012
150
Concentrating solar power technology
Lenzen, M. (1999) Greenhouse gas analysis of solar-thermal electricity generation.
Solar Energy 65(6), 353–368.
Leontief, W. (1966) Input–Output Economics. New York: Oxford University Press.
Leontief, W. (1986) Input–Output Economics, 2nd edn. New York: Oxford University Press.
Linares P., J. Leal and R. Sáez (1996) Evaluación de las externalidades de la biomasa
para producción eléctrica. Documentos CIEMAT.
Norton, B. and W. R. Lawson (1996) Full Energy Chain Analysis of Greenhouse Gas
Emissions for Solar Thermal Electric Power Generation Systems, Vienna: IAEA.
Nuño, F. (2008) Bulk Solar Power Generation: CSP and CPV technologies. Cooper
Institute. Available at: www.leonardo-energy.org (accessed March 2010).
Pehnt, M. (2006) Dynamic life cycle assessment (LCA) of renewable energy technologies. Renewable Energy 31, 55–71.
Preiss, P. and R. Friedrich (2009) NEEDS Technical paper 2° 7.2 – RS 1b ‘Report on
the application of the tools for innovative energy technologies’.
Ragwitz, M., W. Schade, B. Breitschopft, R. Walz and N. Helfrich (2009) EmployRES:
The impact of renewable energy policy on economic growth and employment in
the EU. Karlsruhe: Fraunfofer ISI.
Röder A. (1997) Vergleich regenerativer Energiesysteme bez-u¨glich Kosten, Treibhausgasemissionen, Ressourcenver-Press, Cambridge. brauch. Report No. TM-5197-04, Paul-Scherrer-Institut, as cited in Lenzen (1999).
Tegen S. (2006) Comparing Statewide Economic Impacts of New Generation from
Wind, Coal and Natural Gas in Arizona, Colorado and Michigan, NREL/
tp-200-37720.
Ten Raa T. (2005) The Economics of Input–output Analysis. Cambridge: Cambridge
University Press.
Udo de Haes, H.A., G. Finnveden, M. Goekoop, M. Hauschild, E.G. Hertwich,
P. Hofstetter, O. Jolliet, W. Klopffer, W. Krewitt, E. Lindeijer, R. Müller-Wenk,
S.I. Olsen, D.W. Pennington, J. Potting and B. Steen (2002) Life Cycle Impact
Assessment: Striving towards best practice. SETAC.
Vant-Hull, L.L. (1992) Solar thermal electricity: an environmental benign and viable
alternative. Perspectives in Energy 2, 157–166.
Viebahn, P. (2003) SOKRATES – Projekt, Solarthermische Kraftwerkestechnologies
für den Schutz des Erdkimas, A. P.2.2.
Viebahn, P. (2004) INDITEP, Integration of DSG Technology for Electricity Production, WP 4.3, Impact Assessment, Life Cycle Assessment of Construction Materials,
Energy Demand and Emissions of DSG, Final Report.
Viebahn, P., S. Kronshage, F. Trieb and Y. Lechón (2008) NEEDS Project ‘Final
report on technical data, costs, and life cycle inventories of solar thermal power
plants’ Deliverable 12.2 – RS 1a.
Weinrebe, G., M. Böhnke and F. Trieb (1998) Life cycle assessment of an 80 MW
SEGs plant and a 30 MW Phoebus power tower. Proc. Int. Solar Energy Cont.,
ASME.
Whitely, M., A. Zervos, M. Timmer and F. Butera (2004) Meeting the targets and
putting renewables to work. Overview Report for MITRE Project, ALTENER
Programme. Directorate General for Energy and Transport.
© Woodhead Publishing Limited, 2012
6
Linear Fresnel reflector (LFR) technology
D. R. M I L L S, formerly Ausra Inc., Australia
Abstract: A single-axis linear Fresnel reflector (LFR) system is
composed of many long row reflectors that together focus sunlight
images that overlap on an elevated linear tower receiver running parallel
to the reflector rotational tracking axis. This allows a large size array to
be constructed inexpensively using very similar or identical long focal
length (and therefore almost flat) glass mirror elements. The chapter
reviews the historical and recent technical development of LFR
technology. Thermal performance trade-offs and future trends are also
examined.
Key words: Linear Fresnel reflector (LFR), linear Fresnel, concentrating
solar thermal (CST), concentrating solar power (CSP), solar thermal
electricity (STE), solar concentrators.
6.1
Introduction
Historically, most solar thermal electricity systems have used large continuous curvature parabolic troughs. Geometrically, the ideal reflectors to use
with single receivers of solar energy are continuous reflectors of parabolic
or paraboloidal shape. However, at large scale, these become unwieldy and
may require extensive structures to withstand wind loadings. Operations
and maintenance (O&M) for large mirrors can also become a problem,
since such structures can be much taller than maintenance staff, who then
may require tall vehicles or cranes to perform routine cleaning and
maintenance.
Large reflectors can be simulated by small reflector elements distributed
over some suitable (ground or roof) surface. This allows large concentrator
systems to be built up from small elements, avoiding the large structures
and cleaning accessibility problems associated with very large reflectors.
Linear Fresnel reflector (LFR) solar mirrors are analogues of the parabolic
trough mirror, just as central receiver heliostats are analogues of parabolic
dish collectors (Fig. 6.1).
Today’s usual single-axis tracking LFR differs from a parabolic trough in
that the reflector is composed of many long row segments which focus collectively on an elevated long linear tower receiver running parallel to the
reflector rotational axis (Figs 6.1 and 6.2). Unlike parabolic troughs, the
LFR receiver is fixed in space, and the reflectors rotate to maintain focus
153
© Woodhead Publishing Limited, 2012
154
Concentrating solar power technology
6.1 Basic linear Fresnel reflector configuration seen from one end. The
individual reflectors each track the sun by turning about a horizontal
axis normal to the page.
on the fixed receiver. The reflector rows all track through the same number
of degrees during the day as the sun moves but are inclined at different
angles at any one time because they have different positions relative to the
tower target. This allows a very large basic unit made up of identical relatively small reflectors of long focal length, which in turn leads to the ability
to use almost flat, lower cost glass mirror elements.
The LFR approach to concentrating solar power (CSP) is less commercially mature than trough systems. The field is best examined by looking at
the specific activities of the current major commercial initiatives: Areva
Solar, Novatec Solar, Solar Power Group and the smaller process heat
Industrial Solar company. This chapter reviews the historical development
of LFR technology and then examines the business and technical development history of these players. Thermal performance issues and future trends
are also presented.
6.2
Historical background
Linear Fresnel reflector solar collector systems are called ‘Fresnel’ reflectors after the great French optical physicist Augustin-Jean Fresnel, who in
about 1818 discovered that the effect of large lenses can be duplicated using
many small lens components. However, he was long preceded by the famous
polymath Georges-Louis Leclerc, Comte de Buffon. Buffon had previously
performed experiments in 1746 with the first solar heliostat-like reflectors
(Buffon, 1830), which were manually tracked heliostats similar to those in
modern solar central receiver towers (Fig. 6.2), fashioned out of many
pieces of flat glass installed at slight angles to form a distant focus. Buffon
demonstrated on many occasions that such reflectors would ignite wood
and melt metal. Given the precedence of Buffon and the solar ancestry,
© Woodhead Publishing Limited, 2012
Linear Fresnel reflector (LFR) technology
155
6.2 View of an early heliostat used for experiments by Buffon, who
used many small flat glass mirrors to form a single image in 1746.
Fresnel solar reflector systems might more justly have been called ‘Linear
Buffon’ reflector systems, but the Fresnel name is now too established.
Buffon’s heliostats, however, were not single-axis tracking systems, but
two-axis.
The first person in modern times to apply this principle in a reasonably
large system for solar collection was the Italian solar pioneer Giovanni
Francia (Francia, 1968) who developed both linear and two-axis tracking
Fresnel reflector systems, and was thus the father of both LFR and central
tower systems. Figure 6.3 shows his first prototype array from 1961, built in
Marseille. His papers provided little in terms of theory or detailed efficiency
results, but showed that elevated temperatures could be reached using such
systems.
In the 1970s Francia worked in the United States primarily on central
receiver power tower systems, but continued to maintain an interest in
© Woodhead Publishing Limited, 2012
156
Concentrating solar power technology
6.3 The first LFR prototype set up in Marseille by Francia in 1961
(Francia, 1968) (reproduced with permission from Elsevier).
linear systems and Fig. 6.4 shows a drawing from an Ansaldo/Cesen brochure from about the time of his death in 1980 with his impressions of a
future LFR plant, which show a remarkable similarity to the initial threeline, ten reflector per tower Kimberlina plant built by Ausra Inc. (now
Areva Solar) in California in 2008.
At least three other LFR systems were under development in the 1970s
after the oil shock in 1973. The first was designed by Suntech (no relation
to the current PV manufacturer) and consisted of ten long slightly curved
reflector mirrors 6.1 × 0.3 m, yielding a concentration factor of 40 times. In
a 1970s report, it is mentioned that Sheldahl received a government contract
to develop the concept (USOTA, 1978), but nothing further is known about
the project.
A second design, called ‘Itek’, is shown in Fig. 6.5. This used seven
reflectors and a cylindrical glass receiver containing a restricted aperture,
an absorber pipe and insulation. This was evaluated in 1979 by Shaner
and Duff (1979), who decided that trough collectors had superior
performance.
A third and major effort was described by Di Canio et al. (1979) of the
FMC Corporation, who produced a detailed project design study in 1979
for a linear plant of between 10 MW(e) and 100 MW(e), with a mirror field
on one side of a 1.68 km linear cavity absorber mounted on 61 m high
© Woodhead Publishing Limited, 2012
Linear Fresnel reflector (LFR) technology
157
6.4 Sketch of a LFR solar plant in a desert environment (courtesy of
Cesare Silvi, Italian Group for the History of Solar Energy (GSES) and
the Italian Committee ‘The History of Solar Energy’ (CONASES)).
Receiver
Pyrex tube enclosure
(6 inch diameter)
Insulation
(rein-k)
Movable
mirrors
Absorber rod
(3/4 inch diameter,
39 inches long)
Stainless steel
radiation barrier
Heat trap louvers,
(end louvers reflective
center louver clear glass)
Tracking
Absorber aperture window
mechanism
(1.5 inches wide)
6.5 The Itek LFR concept from the 1970s (USOTA, 1978).
towers adapted from transmission towers. Funding was cancelled just as
initial absorber testing was underway. Unusually, the FMC design was an
east-west axis plant and had an aperture cover that could open and close
to retain heat when the sun was not available.
During the developmental flurry of the 1970s, substantial advances were
made in the areas of spectrally selective absorbers and secondary concentrators, both of which act to alleviate thermal losses from these linear
designs. These were probably the major area where these early LFR systems
© Woodhead Publishing Limited, 2012
158
Concentrating solar power technology
suffered most in the 1970s when compared with higher cost central receivers or trough systems. There is no doubt that there is a link back to the early
experiments of Francia, the founder of this technical area. Experiments
were begun in the USA in the 1970s during Francia’s time spent there, but
after the election of the Reagan Administration, LFR fortunes, along with
CSP generally, fell and interest in LFR technology moved elsewhere.
A new effort to produce a tracking linear Fresnel reflector was made
by the Israeli company Paz which began design work in 1986 and was later
assisted by the Jacob Blaustein Institute at Sede Boqer (Feuermann and
Gordon, 1991; Feuermann, 1993) in southern Israel. Jeff Gordon had
worked with Ari Rabl and Roland Winston, who had published an early
paper on secondary concentrators (Winston and Rabl, 1976) to increase
radiation concentration on a receiver illuminated by a primary mirror.
Intended for 150°C operation, the Paz technology used a secondary concentrator somewhat related to the compound parabolic concentrators first
proposed by Winston and Rabl together with an evacuated tube receiver.
Rows of mirrors were linked together for tracking, and the problems with
linkages, non-parallel mirrors and eccentric mirror mountings led to a
beam uncertainty of more than 2°. Unfortunately, the array exhibited aberration difficulties caused by the movement of reflectors around an axis
parallel to, but displaced from, the reflector optical axis. Performance was
less than 50% of that predicted, but the researchers’ report provided valuable practical lessons for later workers. It was the first to use non-imaging
optical theory to design a secondary reflector near the receiver to increase
concentration, and the first LFR to use an evacuated tube. No photos were
available of the installation but the author of this chapter recalls seeing it
in the mid 1990s. Figure 6.6 is created from drawings in Feuermann and
Gordon (1991).
Before 1993, the LFR concept had each field of reflectors directed to a
single tower. However, if one assumes that the size of the field will be large,
as it must be in technology supplying electricity in the multi-megawatt class,
it is reasonable to assume that there will be many towers in the system. If
they are close enough, then some of the field reflectors will have the option
of directing reflected solar radiation to at least two towers rather than just
one. Development began on an Australian design at the University of
Sydney in 1993, in which a single field of reflectors could use multiple linear
receivers by allowing reflectors to change their focal point from one receiver
to another during the day in order to minimize shading in the dense reflector field. This additional variable in reflector orientation provides the means
for much more densely packed arrays, because patterns of alternating
reflector orientation can be set up such that closely packed reflectors can
be positioned without shading and blocking. The interleaving of mirrors
between two receiving towers is shown in Fig. 6.7. The arrangement mini-
© Woodhead Publishing Limited, 2012
Linear Fresnel reflector (LFR) technology
Metal
tube
159
Glass
tube
Aluminized
reflector
6.6 Drawing of the Paz LFR reflector and tube adapted from
Feuermann and Gordon (1991), reproduced with permission.
6.7 Early 1990s diagram by the author of a CLFR showing interleaving
of mirrors minimizing shading between mirrors.
mizes beam blocking by adjacent reflectors and allows high reflector densities and low absorber tower heights to be used.
The proposed systems were called compact linear Fresnel reflector
(CLFR) systems and their mirrors occupy an area approximately 70% of
the ground area spanned, compared with about 33% for trough fields. The
original CLFR concept was developed with the use of vertically mounted
evacuated tube receivers in mind (Mills, 1995a), later also described in a
paper in 1999 (Mills and Morrison, 1999), but both these and inverted cavity
receivers were anticipated in the original patent (Mills, 1995b).
The CLFR was first developed in ignorance of previous LFR work. It
was conceived of as a multiple line technology from the start, emerging
from a study of minimum reflector area for light capture. However, the
previous work of Francia soon came to light. Later, the Israeli LFR work
came to the attention of David Mills, leader of the University of Sydney
effort, in 1994 and Feuermann kindly supplied Mills with a detailed laboratory report. Although not a CLFR project, the Paz project work provided
the University of Sydney with useful heliostat design information, especially
approaches that did not work well, and this helped define the direction of
field design of the later projects.
© Woodhead Publishing Limited, 2012
160
Concentrating solar power technology
Although a specialist in both non-imaging optics and evacuated tubes,
Mills felt that the hurdles of secondary reflector degradation under high
solar flux, loss of optical efficiency in the secondary reflector (>10% in the
reflectors available at the time) and the high cost of parabolic trough-type
evacuated tubes were serious practical issues. It was decided not to use a
secondary reflector, and instead use a receiver composed either of an array
of proven small low-cost evacuated tubes as developed by the University
of Sydney, or a simple non-evacuated absorber composed of steel tubes with
selective paint. This was an important decision which ultimately led to
Areva/Ausra technology differing from other designs 10–15 years later.
However, the university would not protect the technology, so ownership
of the intellectual property (IP) was re-sold to Mills, and further IP filings
and modelling development were undertaken through Solsearch Pty Ltd, a
company created by Mills and Graham Morrison of the University of New
South Wales. A second CLFR patent covering fully ganged reflector systems
was filed in 1997, but the main research effort was directed at systems
without linked rows because of mechanical issues with linkages. Solar industry partner Solahart also assisted by designing a four-mirror heliostat (Fig.
6.8). After the Kyoto climate meeting, interest increased in CLFR within
the Australian utility industry. Austa Energy in Queensland (not to be confused with the later Ausra LFR company) agreed to develop a 4 MW(e)
CLFR plant with Solsearch under a grant from the Australian Greenhouse
Renewable Energy Showcase scheme in 1999. The project sought to supply
an existing coal-fired plant called Stanwell with extra energy. Mills and
Chris Dey presented a conceptual work on supplementation of coal-fired
plants with LFR technology (Mills and Dey, 1999); nowadays these are
called ‘solar booster plants’ and have become an emerging solar market
sector.
6.8 LFR test rig built by Solahart for the University of Sydney test
programme comparing evacuated and non-evacuated absorbers
(supplied by the author).
© Woodhead Publishing Limited, 2012
Linear Fresnel reflector (LFR) technology
161
6.9 Ray trace illustration of a CLFR array with multiple cavity receivers
in late 1990s (with thanks to University of Sydney School of
Architecture).
By 1999, the Austa/Solsearch project group were actively developing a
cavity receiver design after concluding Solsearch funded experiments at the
University of Sydney, comparing evacuated tubes against side-by-side
blackened steel boiler tubes in a cavity receiver. Single vacuum absorber
tubes were too small to be used as an entire absorber, and multiple small
vacuum tubes exhibited large optical losses in the gaps between the cover
glass and absorber that brought performance down below the cavity (Dey
et al., 2000). New collector designs based on using downward-facing cavity
receivers were researched (Fig. 6.9) and visualized, and flow modelling was
undertaken (Reynolds et al., 2001). The receiver was envisaged as an
inverted, trapezoidal, linear cavity receiver with a window at the base of
the cavity that was larger than the receiver. The insulated cavity was trapezoidal to allow concentrated light from the reflector field to strike the
absorber directly without incurring an optical loss in a secondary reflector.
The ‘window’ had a transmissivity of 0.95 at solar wavelengths. The absorber
was envisaged as either a flat plate attached to tubes or tubes bonded
together. In addition, an initial heliostat prototype was developed by the
industrial collaborator, Solahart (Fig. 6.10). This had a ‘backbone and rib’
structure.
In 1993, Belgian investors bought the bankrupted Luz technology assets
for trough collectors and evacuated tubes. They created two companies,
Solel Israel and Solel Europe, based in Belgium. In 1994, these investors
entered into a commercial in-confidence agreement with the University of
Sydney regarding the early CLFR technology. Concepts such as nylon gears
running flat mirrors with torque tubes, long single one-ended steam receiver
pipes with internal water feed tubes, and non-imaging CPC-type secondary
reflectors without vacuum receivers were all discussed in confidence at the
© Woodhead Publishing Limited, 2012
162
Concentrating solar power technology
6.10 Prototype backbone and rib heliostat produced by Solahart in
Perth in 1999 for the project (courtesy of Solahart Pty Ltd).
time, but the joint venture was discontinued by the university when Solel
Europe became insolvent in 1995. A detailed ‘recipe’ for the revolutionary
evacuated tube double cermet solar selective absorber coatings (Zhang and
Mills, 1992) by the University of Sydney was passed in confidence to Solel
Israel about two years ahead of the announcement of the UVAC2 receiver
in 1997. Israeli staff stated to the author they had been close to patenting
a similar surface when the Sydney University patent was filed in 1991. The
university selective surface coating is now used on a very large scale in
China for evacuated tube hot water heaters.
In 1998, one of the former owners of Solel Europe co-founded the
Solarmundo company using private funds from Spain (Manuel Sureda) and
Belgium (Count de Lalaing). By 2001 Solarmundo had built a 2,400 m2
prototype collector field of the LFR type at Liège in Belgium (see Fig. 6.18).
Delays in the Australian project due to the breakup of the Queensland state
utility, Austa Energy, ensured that the Solarmundo array became the first
project of any size to be demonstrated since the work of Paz a decade
earlier, and it was the largest LFR array ever built up to that time. While
the prototype used some concepts discussed in Sydney in 1994/5, ‘mirror
flipping’ was not used, so it was not a CLFR.
The following sections follow the four major commercial initiatives currently offering LFR systems. Names of all participants have changed, so the
current company name is used to head the sections with the previous names
in parentheses. It is important to note that the assessment of any particular
system is based on the published data available. Comparisons are sometimes difficult to make without access to what continues to be proprietary
commercial information. Where it has been necessary to make any assumptions in assessing a particular system, these have been clearly stated so that
readers can reach their own conclusions. Where there is insufficient published data available on a system to make an assessment, it has not been
discussed in this chapter.
© Woodhead Publishing Limited, 2012
Linear Fresnel reflector (LFR) technology
163
6.11 First sketch of the SHP design from early 2002 (courtesy of
P. Le Lièvre).
6.3
Areva Solar (formerly Ausra, Solar Heat
and Power)
In late 2001, Mills, Morrison and Peter Le Lièvre formed a new company
in Sydney called Solar Heat and Power (SHP). Previous IP from Solsearch
and the University of Sydney efforts was taken over by the new company.
Le Lièvre designed a support structure for the reflectors based upon a space
frame concept that is still used by Areva Solar today.1 He also worked with
the other company founders to develop the remainder of the design, which
was built upon Solsearch know-how and IP. Figure 6.11 shows a sketch of
the first small array that was planned at the time.
In keeping with the Solsearch design philosophy, secondary reflectors
were not used, partly to avoid losses in optical efficiency in mirror absorption, and partly to avoid rapid degradation of the reflector surfaces under
high heat loading. A new tracking design using low-cost drive hoops was
chosen to turn the reflectors. The previous Solsearch/University of Sydney
work had proposed trapezoidal receivers with welded or thermally connected steel tubes, but the new SHP design separated the tubes in the
receiver, allowing differential thermal expansion between tubes at slightly
different temperatures. This was analysed for thermal convective and radiation losses in a paper by Pye et al. (2003). Figure 6.12 shows a schematic
diagram from the paper representing a trapezoidal receiver.
In 2003, an agreement with Macquarie Generation in New South Wales
was concluded to develop the technology at Liddell Power Plant as a coal
plant booster. By 2004, SHP had built a 61 m long 1,340 m2 prototype LFR
on the grounds of the Liddell coal-fired power plant near Singleton in New
1
Dr David Mills has not worked with AREVA Solar since June 2010 and has no direct knowledge of Areva Solar’s present-day technology.
© Woodhead Publishing Limited, 2012
164
Concentrating solar power technology
Absorber surface (steel pipes)
Insulation
Sidewalls
Radiative
losses
Convective
losses
Cavity cover (plastic film)
6.12 Schematic of a trapezoidal inverted cavity receiver with a plastic
transparent cover. Ultimately, a glass cover proved to be more
durable (courtesy of G. Morrison).
6.13 Stage 1 of the Liddell array in 2004 soon after initial operation
(copyright Areva Solar, used with permission).
South Wales that was smaller than the 2,400 m2 array erected in 2001 by
Solamundo. This was a 1 MW prototype (Fig. 6.13) not connected to the
power block and intended only to demonstrate operation at the temperatures and pressures required by a power plant preheater (280°C and 80 bar).
The SHP prototype achieved its design goals quickly, and later was operated
briefly to demonstrate ‘once through’ production of superheated steam.
Figure 6.14 shows the flow schematic of the prototype. A journal paper was
published on the initial prototype and future plans in 2006 (Mills et al.,
2006). The absorber in the first SHP prototype consisted of 16 parallel DN
25 pipes, each 60 m long, made of 304 stainless steel, and mounted
© Woodhead Publishing Limited, 2012
Linear Fresnel reflector (LFR) technology
Absorber tubes (120 L capacity)
Pressure relief
P∝ at 260°C
Level
Steam switches
drum
100 MPa
450 L
Circulating
pump
165
Inlet water tank
Feedwater
pump
6.14 Schematic of prototype system at Liddell, which was not
connected to the coal-fired station (courtesy of G. Morrison).
side-by-side with very little spacing, for a total absorber width of 575 mm.
Absorber pipes in the prototype were connected via a set of manifolds into
a four-pass configuration, so that the total length of a flow path was approximately 240 m, emulating a much longer collector. Water was introduced into
the outer tubes in a preheating arrangement and returned via the centre
tube, where it turned into saturated or superheated steam. The system thus
had connections at one end and the other end was free to move to accommodate thermal expansion. The tubes were suspended by rollers positioned
periodically down their length. Anti-reflection layers were introduced on
the inverted cavity glazing. The tubes were coated with black non-selective
coating, awaiting commercial deployment of an in-house developed selective coating.
In 2004 Mills delivered papers suggesting that very large plants using
LFR technology around 300°C could use low pressure nuclear type turbines
as a low-cost generation possibility (Mills et al., 2004a, 2004b). At the same
conference, a study in Germany (Häberle et al., 2004) was presented suggesting that LFRs were potentially viable as a competitor to troughs. At
this time there were no high temperature air stable selective coatings, so
operation at high efficiency would be limited to that of air stable coatings
like Black Chrome, which is stable up to about 300°C. This temperature was
only slightly above the preheating temperatures that would be required for
coal plant booster arrays discussed by Mills and Dey in 1999, including the
next stages of the Liddell project.
In 2006 the Liddell project began Stage 2, a 20,000 m2, 5 MWe stage that
was connected to the main station power block to supply thermal energy
to the final feedwater heater. Stage 2 used an improved commercial model
of the collector (Fig. 6.15) with several changes from the first prototype. The
mirror reflectivity was increased from 84% to 92.5%, the reflector width
increased from 1.82 m to 2.25 m, the length increased from 12.2 to 12.9 m
© Woodhead Publishing Limited, 2012
166
Concentrating solar power technology
6.15 Stage 2 of the Liddell array in 2006 (copyright Areva Solar, used
with permission).
and the number of reflectors per absorber line reduced to ten from the
original twelve. The tower and absorber dimensions were unchanged.
In late 2006, SHP was invited to start up a US operation funded by Silicon
Valley venture capitalist companies, named Ausra Inc. (‘Ausra’ is one
version of the ancient Indo-European word for the Goddess of the Dawn),
ultimately obtaining US$130 million in international venture capital. Ausra
subsequently bought all SHP assets including all IP, and the centre of
technical operations shifted to California. SHP became Ausra Pty Ltd in
Australia.
In 2008, the company built a redesigned pre-commercial unit of three
384 m lines located at Kimberlina in Bakersfield, California (Fig. 6.16). This
employed a redesigned A-frame tower (similar to the 1970s Itek tower and
the contemporary Industrial Solar design) to increase earthquake resistance and simplify receiver installation and maintenance. The unit was
connected to a dedicated turbine, a 5 MW former biomass steam turbine
that was already installed on the site for a previous project. The plant was
the first LFR built for electricity production in North America and was
opened in October 2008 by Arnold Schwarzenegger, the Governor of
California, who noted it was the first solar thermal electricity plant of any
kind to enter operation in California in nearly 20 years. Unfortunately,
serious technical problems arose with the second-hand turbine that were
unrelated to the collector system, and these were not repaired until the
spring of 2009, when the installation was finally connected to the California
grid.
The field length at Kimberlina is given by NREL as 385 m (Ausra, 2007).
In the first three lines built there were 25,988 m2 in 30 reflector rows each
© Woodhead Publishing Limited, 2012
Linear Fresnel reflector (LFR) technology
167
6.16 The three-line Kimberlina array in Bakersfield, California, in late
2008 of LFR technology (copyright Areva Solar, used with permission).
having 24 modules per line. Thus, each mirror module has close to 36.1 m2
of reflector in five reflector panels 2.25 m wide, for a module length of
16.0 m. The mirror modules used at Kimberlina were manufactured at
Ausra’s automated solar thermal power factory in Las Vegas, Nevada,
which is designed to produce one reflector module every 8 minutes, or
4 m2 per minute, or 3,360 m2 over two 7-hour shifts per day. The 2008 array
was designed for 300°C saturated steam, but was also used as a test unit
for superheated operation throughout 2009. Kimberlina has since been
used intensively to verify the array performance for prospective customers
and investors.
However, for large projects the company needed significant backing.
By 2010, competitors like Solel, Solargenix, Solar Reserve and SES had
already been purchased or backed by much larger companies. In early
2010, Ausra was bought outright by the French nuclear giant Areva and
was renamed Areva Solar. Areva is experienced in mounting large projects,
so the company is now effectively in the business of selling both steam
technology and power projects. In early 2010, a $105 million 18 line
44 MWp(e) steam booster project was secured with the utility CS in
Queensland to go online in 2013. The system will operate at 330°C. Its
technology has been selected for a 250 MWp(e) flagship solar project in
Queensland (Solar Dawn, 2011), but this was shelved in July 2012 due to
lack of a buyer for the power. However, a similar 250 MW plant announced
for India in 2012 is under construction by AREVA at the time of writing.
In 2011 Areva Solar also announced a memorandum of understanding to
commence engineering studies for a 150 MW CLFR free-standing plant
(PEM, 2011) to be installed near Fresno, California.
© Woodhead Publishing Limited, 2012
168
Concentrating solar power technology
6.17 The fourth line at Kimberlina in operation. It differs from the first
three lines in using 13 reflectors instead of ten, and in operating as a
superheating line at 400°C (copyright Areva Solar, used with
permission).
Areva Solar has more recently tested a fourth, newer technology line
called SSG4 (Conlon, 2011) at Kimberlina in 2010, pressurized to 92 bar
(Fig. 6.17) that uses a 13 mirror line configuration with a superheated steam
receiver, unlike the previous ten mirror line saturated steam units. In the
fourth line there are 11,261 m2 in 13 reflector rows each having 24 modules
per line. We previously learned each module is 16.0 m long and the reflectors 2.25 m wide, yielding 29.25 m2 of reflector per lineal metre of receiver.
Areva states the heating surface area to be 210.0 m2 for a line, so the
heating surface must be 210/385 = 0.545 m wide. The peak ratio of primary
reflector to heating surface area for the smaller 545 mm wide receiver is
thus about 29.35/.545 = 53.7. The performance tests below 400°C were
overseen by the consultant engineers Black and Veitch and 100% availability was achieved during the June–October testing period in 2010. Performance is said by the company to have met or exceeded modelling
predictions (Conlon, 2011). Interestingly, Conlon says that in a ‘lights out’
test, the array line had enough thermal inertia to deliver 18 minutes of
superheated steam.
Areva has stated that it is developing a 482°C version available for 2011
with 165 bar operation (Areva, 2012) , which may be the 13 reflector line
version described by Conlon (2011), who cites a design operational temperature of 450°C and pressure of 165 bar with a maximum pipe wall temperature of 482°C. Areva has also suggested a version 2.0 to be under
development, so it is possible that the next version could be designed for
maximum steam temperature of more than 500°C, but that is not clear at
the time of writing.
© Woodhead Publishing Limited, 2012
Linear Fresnel reflector (LFR) technology
169
6.18 The Solarmundo prototype at Liège (copyright Solarmundo
2001).
6.4
Solar Power Group (formerly Solarmundo,
Solel Europe)
As stated in Section 6.2, the founder of Solel Europe joined forces with
Belgian investors to found Solarmundo in the late 1990s. The design was
an LFR using a non-imaging receiver secondary reflector for use with a
custom-built non-evacuated absorber tube, a similar concept to the earlier
Paz receiver design.
The Liège prototype (Fig. 6.18) built by Solarmundo was operated as a
test bed until 2004 when the company was closed down. The Count de
Lalaing then founded Solar Power Group (SPG) to continue his work on
the LFR. This prototype marked the establishment of German interest in
the technology; before the move to Germany, Solarmundo used the
resources of German research institute PSE to attack the issues of optical
reflector design, selective coating durability at higher temperatures, and
secondary mirror stability. In 2005, SPG entered a cooperative agreement
with Ferrostaal, a subsidiary of the industrial group MAN. In 2007,
Ferrostaal acquired 25% of SPG, and later in 2009, increased its ownership
to 43%.
Because of generous German government assistance, the majority of
publicly available research on LFR work has for a long period been associated with the SPG technology, so that general conclusions about LFR
viability have tended to use the SPG technology as a quasi-standard, rather
than the more secretive but commercially active Areva and Novatec Solar
technologies. A great deal of early research has emerged into the public
© Woodhead Publishing Limited, 2012
170
Concentrating solar power technology
6.19 The FRESDEMO SPG prototype at PSA. The tower has been
made narrower in the newer version and bracing is less optically
intrusive than the earlier prototype (copyright MAN Ferrostaal and
Solar Power Group GmbH 2007).
sphere because of the SPG project (Bernhard et al., 2008; Morin et al., 2006;
Mertins et al., 2004; Eck et al., 2007).
In 2007, the company undertook construction and operation of their
FRESDEMO pilot collector in the Plataforma Solar de Almería (PSA) in
Spain (Figs 6.19 and 6.20). Land for its construction was provided by
CIEMAT at the PSA. The DLR’s Institute of Technical Thermodynamics
analysed the thermal characteristics and thermal losses of the linear Fresnel
collector in order to determine the thermal collector efficiency at different
receiver temperatures. Fraunhofer ISE was assigned by MAN to validate
the collector models of the optical efficiency and thermal losses with the
measurements.
The FRESDEMO re-optimized design had a length of 100 m and a total
width of 21 m including 15 m of cumulative mirror surface width. The
absorber tube dimensions consist of a 14 cm outer diameter with an approximate 12.5 cm inner diameter. There are 15 m2 of reflector per linear metre
of absorber tube, so the geometrical concentration (the ratio of aperture
area to receiver emitting surface area) is 34, about 20% higher than
© Woodhead Publishing Limited, 2012
Linear Fresnel reflector (LFR) technology
171
6.20 Public showing of the FRESDEMO prototype at the PSA in Spain
(copyright MAN Ferrostaal and Solar Power Group GmbH 2007).
parabolic troughs. The steel structure of the prototype supports a fixed
central absorber tube located in the centre of a secondary reflector and 25
rows of slightly curved primary mirrors.
The FRESDEMO collector can be operated in three different operational modes that can be pre-selected:
•
•
•
preheating (absorber fed with cold or preheated water)
evaporation (absorber fed with preheated water and saturated steam)
superheating (absorber fed with saturated or superheated steam only).
Inside the cavity is a single absorber tube with an inner diameter of 18 cm
and coated with a non-selective absorber coating. The optical efficiency
measured at PSA (Bernhard et al., 2009) first suggested 63%, and this efficiency reduced to about 53% over some months. However, periodic calibrations of the primary mirrors, as recommended by PSE AG, were not
performed for various reasons. After recalibration and a complete cleaning,
the original efficiency of over 60% was achieved again in 2009. Also the
soiling of the glass plate below the cavity could account for about 2% efficiency decrease per month and measurements of glass plate soiling did not
commence until June 2008, nearly 30 days after initial cleaning, so a peak
optical efficiency of 64% is likely to have been the initial state. A raytrace
calculation suggested a peak optical efficiency of 65.2%, close to the 64%
estimated from measurements.
The FRESDEMO design lost about 850 W of thermal energy per metre
of receiver length at 300°C according to Bernhard et al., and has 15 m2 of
reflector per lineal metre of receiver length, so the thermal loss was about
57 W/m2 of field reflector, higher than Novatec Solar and Industrial Solar
© Woodhead Publishing Limited, 2012
172
Concentrating solar power technology
contemporary secondary reflector designs as calculated later in this chapter
(see Fig. 6.29). This is likely to be because the Novatec design has significantly higher optical concentration on its receiver, while the Industrial Solar
design uses a low loss evacuated tube receiver. However, that information
was published relatively recently. Before that, early low temperature testing
<300°C by all LFR manufacturers had fostered a general assumption in the
solar industry that LFRs were low temperature devices by nature, an
opinion only now being abandoned as most new model LFRs are now being
designed and tested at or above 450°C, higher than most current PTCs
(parabolic trough collectors). FRESDEMO finally broke the myth in September 2008, operating at up to 450°C an in-line superheater for a trough
direct steam generator (SPG, 2008), the highest LFR temperature so far
achieved at the time.
Cooperation with German technology groups has led to significant progress in the reflector optical design and sputtered selective coating area.
Selective absorber surfaces have been developed that may be reliable at
450–500°C, though the cost of these somewhat complicated layered sputtered surfaces is not known. There has been relatively little available
description of this surface work in the open literature, but the SPG Managing Director Count Jacques de Lalaing described the effort as follows in a
radio interview (de Lalaing, 2009):
[O]ne of the main differences that we have with our competitors is that we’ve
developed a coating that is holding at a higher temperature and it allows us to
reach 450 degrees . . . we can produce steam at 450 degrees in a very stable
condition, super-heated steam, and of course that gives us an edge over the
competitors because the higher the temperature of your steam, the higher the
efficiency of the turbine that you hook up behind it.
The high temperature mirror stability issue has been more difficult, and
progress was slower, but the problem has been taken up by glass reflector
manufacturers who are expected to offer products soon that will allow
high reflectance secondary reflectors suitable for high temperature cavity
receivers.
Within the extensive paper by Bernhard et al. (2008), researched by SPG
together with the Fraunhofer Institut für Solare Energiesysteme (ISE) and
the German Aerospace Center (DLR), the technical aims of SPG are
clearly stated:
SPG is planning a future collector design in which the primary mirror accuracy
is increased significantly. A new substructure for the primary panels is being
developed together with an automated gluing process and an integrated optical
measurement system to supervise the manufacturing process. This will guarantee long lasting and stable primary mirrors with higher accuracy, manufactured
in a mass production process. Quality management systems will be integrated
© Woodhead Publishing Limited, 2012
Linear Fresnel reflector (LFR) technology
173
also in the assembly process into the steel structure. By using a more accurate
sun tracking algorithm and drives with a higher accuracy as well as inclination
angle encoders with an acute sampling rate, higher precision resulting in a
considerably higher optical efficiency will be achieved. The revised receiver
design will omit openings to the atmosphere preventing efficiency losses due
to rain water and dust ingress.
It is interesting to note that SPG does not, in the above, mention the use
of evacuated tubes. It may be that they are undecided as to whether there
is a value proposition in their use, or have decided not to use them, or are
not disclosing an evacuated tube programme. Their design would require
a single evacuated tube receiver larger than current evacuated tubes. Morin
et al. (2009) have suggested that this substitution of a 70 mm evacuated
tube would substantially reduce heat loss and improve overall efficiency,
even though annual optical efficiency would drop by 10% because the
70 mm receiver is half the original absorber size. The collector and the
Schott receiver have not been designed for each other, and a situation
where the collector is optimized for the receiver would be more optimal.
In spite of this, Morin et al. came to the conclusion that even with too
small an evacuated tube, the advanced model LFR with the smaller tube
had a break-even cost of 80% of the PTC cost, while the current model
breaks even at 53% of the PTC cost. This appears, at first glance, to be a
strong argument for changing to evacuated tubes in an extensively redesigned collector if the financial assumptions are correct. SPG’s competitor
Novatec is doing just that. Also, higher optical concentration is not mentioned, although this would be preferable. The ratios of primary reflector
area to hot surface area achieved by Novatec Solar and Areva Solar are
above 50, but for SPG only 34. It would be surprising if this were not
increased in SPG’s next model.
In terms of project development, SPG has had a long-standing interest
in the Middle East and North Africa (MENA) market, and signed a memorandum of understanding (MOU) with Libya in 2006. This is explained on
their website, which states ‘Of special interest to SPG and MAN Ferrostaal
is the MENA region, that not just offers optimal climatic conditions but
also is the home of IPIC (International Investment Petroleum Company),
which has recently acquired a 70% participation in MAN Ferrostaal’.
In 2010, SPG signed an agreement with GDF Suez for the construction
a 5 MW(th) add-on onto a coal-fired power plant in Mejillones. The construction of the plant should start in 2011 and will use the new design
developed by SPG over the last two years. In early 2011 SPG also signed a
licence agreement with JFEE, a large Japanese EPC company, that will use
SPG technology in plants to be built in southeast Asia and Oceania. It states
(SPG, 2011) that it is ‘developing projects in parallel in Southern Europe,
Northern Africa, South America, and Australia, varying in scale’.
© Woodhead Publishing Limited, 2012
174
Concentrating solar power technology
6.5
Industrial Solar (formerly Mirroxx, PSE)
PSE AG is a solar energy technology and service company that was a spinoff of Fraunhofer ISE, the noted solar research institute in Freiburg. PSE,
under the leadership of Andreas Häberle, assisted Solarmundo from 2000
onwards in the area of optical design, collector performance modelling and
selective surface development, and much of the research that assisted SPG
also came from this source. However, PSE developed a small LFR of its
own for rooftop mounting using a single 70 mm diameter Schott evacuated
tube receiver by 2005. Intended for the process heat market, it employs an
A-frame support similar to the previous Itek collector from the 1970s and
the later Areva system. The system, called the FL-11, uses a water circuit
pressurized at 16 bar to transfer process heat with temperatures up to
200°C. This falls within the pressure restrictions of the evacuated tubes used,
which were designed for the pressures used in oil heat transfer systems.
The basic reflector system has modules that are 4 m long × 8 m wide with
11 primary mirror rows. Like other companies, it uses flat white glass
mirrors, possibly slightly curved elastically, and a polished aluminium secondary reflector. Each of the individually driven mirror rows features an
electric drive motor.
A second prototype with an aperture area of 132 m2 and output
66 kWp(th) was installed in Bergamo, Italy to power an ammonia-water
absorption chiller. It was operated and monitored from August 2006. In late
2007, a third Fresnel process heat collector with a 352 m2 aperture area
attaining a peak of 176 kW(th) was installed on the roof of the Escuela
Técnica Superior de Ingenieros (ESI), a School of Engineering building in
Seville, Spain (Fig. 6.21). The collector total length is given as 64 m (16 ×
4 m long modules) and otherwise similar in design to the ones in Freiburg
and Bergamo. Each module thus has 5.5 m2 of reflector per lineal metre of
receiver, and 0.5 m per reflector per lineal metre of receiver. As each reflector including spacing is 4.06 m, a 4 m reflector length is likely and the
reflector width is about 2 m2 in area. The collector powers a double effect
H2O/LiBr absorption chiller with a maximum cooling capacity of 174 kWth
for air-conditioning the building. At this site, the wet cooling tower for heat
rejection, which is usually necessary for H2O/LiBr absorption chillers, was
substituted by a water heat exchanger fed by water from a local river. The
double-effect absorption chiller offers a coefficient of performance (COP)
of up to 1.3. A fourth project was a solar cooling system with a NH3/H2O
chiller at a winery in Tunisia.
The Mirroxx company was launched as a PSE spin-off in December 2008,
taking over PSE’s LFR commercialization activities. With Mirroxx GmbH,
a basis for industrial series production and strategic marketing of the
Fresnel collector technology had been formed. In November 2010, Mirroxx
© Woodhead Publishing Limited, 2012
Linear Fresnel reflector (LFR) technology
175
6.21 The third Industrial Solar project, built in 2007 in Seville, Spain,
with an aperture area of 352 m2, length 65 m, pressurized water circuit
at 16 bar, and operating temperature 180°C. The application was solar
cooling with a double-effect absorption chiller (175 kW) (courtesy
Industrial Solar GmbH, reproduced with permission).
6.22 Industrial Solar Fresnel collector field in Doha (Qatar) (built in
2010, aperture area 1,408 m2, length 65 m, pressurized water circuit at
16 bar, operating temperature 180°C, application: solar cooling with
double-effect absorption chiller (750 kW)) (courtesy Industrial Solar
GmbH, reproduced with permission).
commissioned its largest collector field so far in Doha (Qatar) (Fig. 6.22).
The heat generated by the collector field is used to power an absorption
chiller, which provides air-conditioning for a 500-seat showcase football
stadium constructed by ES-Group/London and designed by Arup Associates. The system was described in the December 2010 issue of Renewable
Energy World (Appleyard, 2010). The solar thermal field, supplied by
German engineering and manufacturing group Mirroxx GmbH, features
single-axis tracking flat-plate mirrors which focus solar energy onto a Schott
PTR® evacuated tube receiver using water as the heat transfer fluid, pressurized to 16 bar at 200°C. Collector aperture area is 1,400 m2, peak thermal
output 700 kW, and it was stated that Mirroxx claimed a maximum optical
© Woodhead Publishing Limited, 2012
176
Concentrating solar power technology
efficiency for direct normal irradiation (DNI) of 62%, slightly lower than
the more recent figure below.
In April 2011, Mirroxx changed its name to Industrial Solar ‘to reflect the
company’s focus on industrial applications’ (Industrial Solar, 2011a). The
target market consists of industries in sunny countries which require process
heat up to 400°C. In essence, the company has been steadily improving the
original Israeli Paz system secondary reflector concept also favoured by
SPG and Novatec Solar to the point where it has the basic requirements
for not only an industrial system, but a high temperature superheated LFR
steam generation system, including a selectively coated evacuated tube, and
a secondary high efficiency reflector. At 400°C, Industrial Solar is delivering
thermal energy in the same temperature range as the electricity LFR companies, and could therefore conceivably be a player in the high temperature
electricity generation sector in the future, although the company has not
announced any such ambitions. In the downloadable 2011 Industrial Solar
brochure, it is stated that ‘it is also possible to directly generate or even
superheat steam.’ In addition, it was stated that although 40 bar was now
standard, pressures up to 120 bar were offered, greater than the Novatec
Solar Supernova steam pressure of 90 bar (Paul et al., 2011), even though
Novatec central generation LFR uses evacuated tubes as well. The Industrial Solar brochure also provides useful technical data (Industrial Solar,
2011b) for the FL-11 collector and reference conditions as follows:
• Thermal loss at 400°C (μ = 0.00043 W/(m2K2)
• Reference temperature conditions: 30°C ambient; 160°C inflow; 180°C
outflow
• Angle-independent optical efficiency (with 100% clean primary and
secondary reflectors and receiver glass tube)
• Optical efficiency η0 = 0.635 (for sun in zenith)
• ηmax = 0.663 for sun at 5° transversal zenith angle
• Mirror reflectivity 95%
• Schott PTR®70 Receiver thermal emittance @ 380°C: 9%
• Solar absorptance direct: 95%
6.6
Novatec Solar (formerly Novatec-Biosol,
Turmburg Anlagenbau)
In 2005, a new LFR company was formed, called Turmburg. Its technology
founder, Max Mertins, had previously done modelling work for Fraunhofer
ISE where he worked extensively on the Solarmundo collector. Mertins had
also participated in a joint German government LFR workshop (Morin
et al., 2006) attended by both SHP’s European division and SPG and was
thus familiar with both designs. ‘VDemo-Fresnel’ was funded by the German
© Woodhead Publishing Limited, 2012
Linear Fresnel reflector (LFR) technology
177
6.23 End view of a Novatec reflector showing polymer support
bearings. It is of monocoque construction (courtesy Novatec Solar
GmbH).
Ministry of Environment, Nature Conservation and Nuclear Safety (BMU).
The new company produced a design using a CPC cavity receiver somewhat
resembling the SPG design, but using a 70 mm receiver size like the PSE
design, with a supporting design tower similar to the small ‘V’ support used
in the V1.1 SHP CLFR.
A primary design difference between the Turmburg and other designs
was in reflector construction. The new design was a monocoque one, where
much of the strength is derived from the metal skin (Fig. 6.23). The reflectors are lightweight and can be lifted without the need for machines. They
are amenable to mass production and an automated production unit was
designed and built based on automobile component production methods.
In late 2006, the company received investment and commissioned a prototype plant in Spain as a basic module suitable for 270°C operation
(Fig. 6.24). The name of Novatec-Biosol was adopted for the company. In
this section, the company will be referred to as Novatec Solar, the most
recent name at the time of publication, or ‘Novatec’.
The basic solar boiler module (NOVA-1, 2011a) was called Nova-1 and
has a receiver height of 7.4 m measured from a mirror hub height 1.2 m
above the ground. It uses 128 reflectors in 16 reflector lines, 8 to a line, with
a cumulative width of 12 m (0.75 m individual reflector width) and a total
module reflector area of 513.6 m2. This can be calculated by dividing the
quite precisely provided 18,489.60 m2 in the previous reference for the array
PE-1 by 36, the nearest whole number of modules. This calculates to an
individual reflector area of 4.01 m, and an individual reflector length of
5.35 m. The module length is 44.8 m. Each module thus has 513.6/44.8 =
11.5 m2 of field reflector per lineal metre of receiver. The total
© Woodhead Publishing Limited, 2012
178
Concentrating solar power technology
6.24 The 2005 Novatec demonstration module in the south of Spain
(courtesy Novatec Solar GmbH).
6.25 Illustration of Novatec dry cleaning robots moving down a
reflector line (courtesy Novatec Solar GmbH).
circumference of the receiver is 0.22 m, so the ratio of reflector to hot
receiver surface is about 52. A cleaning robot has been developed to keep
the reflectors at high efficiency (Fig. 6.25). This moves down the line automatically and does not use wet cleaning.
The automated production line started to produce primary mirrors in
May 2008, and the first 4,600 manufactured primary mirrors were installed
in the demonstration plant PE-1, located in Calasparra (Spain), with a
nominal capacity of 1.4 MWe. The two parallel collector rows of PE-1 (see
Fig. 6.26) with a length of 806.4 m produce a steam–water mixture at up to
55 bar (270°C). This is separated into steam and water phases in a steam
© Woodhead Publishing Limited, 2012
Linear Fresnel reflector (LFR) technology
179
6.26 PE-1 1.4 MW powerplant (courtesy Novatec Solar GmbH).
drum with the saturated steam directly feeding the turbine while the water
phase is mixed with the feedwater and recirculated to the inlet of the solar
field (Fig. 6.27).
Alluding to its location Puerto Errado in Spain, PE-1 was connected to
the Spanish electrical grid in 2009, slightly in advance of the Ausra Kimberlina project. A company online brochure (NOVA-1, 2011a) claims an optical
efficiency of 67%, and an operating temperature of 270°C. One year later
it was followed by the start of construction of 30 MW PE-2 with 302,000 m2
of collector area (NOVA-1, 2011a). The latter plant is expected to be completed in March 2012.
Data released thus far (NOVA-1, 2011b) on the Nova-1 technology are:
• Convective thermal loss at coefficient μ0 = 0.056 W/(m2K)
• Radiative thermal loss at coefficient μ1 = 0.000213 W/(m2K2)
• Power lost Ploss = μ0 ΔT + μ1 ΔT2
• Reference temperature conditions: 40°C ambient; 100°C inflow; 270°C
outflow
• Angle-independent optical efficiency η0 = 0.67 (for sun in zenith) (with
100% clean primary and secondary reflectors and receiver glass tube)
• No wind assumption stated
• 246.2 kW per module
• 541 W/m2 collected per area of primary reflectors (502.3 W/m2 PE-1)
• 900 W/m2 direct normal radiation (DNI) at azimuth angle 0°, zenith
angle 30°.
In the Nova-1 projects, Novatec Solar uses the low temperature saturated
steam approach earlier used by SHP, but like Areva Solar and SPG, they
© Woodhead Publishing Limited, 2012
180
Concentrating solar power technology
Steam + water
Saturated steam
2
4
3
Recirculation
5
6
11
10
7
1
Feedwater
9
1. Solar field
2. Steam separator
3. Steam storage
4. Turbine
5. Generator
6. Air-cooled condenser
8
Condensate
7. Deaerator/feedwater tank
8. Feedwater pump
9. Recirculation pump
10. Public electricity grid
11. Moisture separator and reheater
6.27 PE-1 1.4 MWe powerplant schematic. The system uses a
saturated steam turbine. From Novatec PE-1 brochure (courtesy
Novatec Solar GmbH).
are developing superheated systems. In 2010 Novatec announced the development using Schott evacuated tubes instead of the current non-evacuated
tubes (SuperNOVA, 2011). The new collector model was called SuperNOVA and was used at the PE-1 site prior to September 2011. Novatec
state that heat loss is reduced by 50% compared to the Nova technology,
but not at which temperature this takes place. Although the company
announcement brochure announces 500°C, Paul et al. (2011) of Novatec
Solar used a steam temperature of 450°C at 90 bar in their calculations for
a 50 MW(e) Supernova, so this might be the normal output temperature
like Novatec’s competitors, even though higher temperatures are possible.
(450°C is also specified in the Areva Solar SSG4 array with a peak wall
temperature of 482°C at a stamped pressure of 92 bar, and by SPG in
superheating tests of the FRESDEMO.) The increase in electrical output
for a 100% solar plant was calculated by Paul et al., and the annual electric
yield is 0.322 MWh(e)/m2 for a SuperNOVA operating at 450°C and 90 bar,
compared to a NOVA-1 value of 0.298 MWh(e)/m2 operating at 285°C and
70 bar, an 8% improvement. Paul et al. also point out that power block costs
are lower at 450°C than 285°C, and that ‘a further improvement including
© Woodhead Publishing Limited, 2012
Linear Fresnel reflector (LFR) technology
181
a reheat stage in a SuperNOVA power plant – as has been done for NOVA-1
– will allow further efficiency improvement’.
Novatec has won a contract to add to the existing solar facility built by
SHP at Liddell power station in New South Wales, Australia, using Nova-1
technology. The new array will deliver a peak of 9 MW(th) and has a
planned array area of 18,490 m2. It was stated to be expected to commence
in early 2011 and be completed in 2012.
On 16 March 2011, Novatec announced that the large energy technology
company ABB had signed an agreement to buy a 35% shareholding in
Novatec Solar, including an option to acquire 100% of Novatec Solar and
an agreement to cooperate on future solar power plant projects.
6.7
LFR receivers and thermal performance
Thermal efficiency and heat losses are critical to design viability. One might
think that there would be only one optimal design, but the Areva Solar
multiple tube cavity receiver is now becoming significantly divergent from
the SPG, Novatec Solar and Industrial Solar secondary reflector designs.
Historically, the original decision by SHP to avoid hot mirrors and secondary reflector absorption losses led to a very optically efficient receiver
design based on low-cost multiple tubes of small diameter arranged horizontally to maximize the effectiveness of the cavity receiver. While it uses
inexpensive reflectors to the side of the cavity to capture some stray light,
these reflectors stay out of the main beam; the receiver does not use secondary reflectors as a primary method of light capture. However, this system
requires the development of steam control systems for very long multiple
tube arrays, a technical problem at the leading edge of steam engineering
practice. The company has been also developing a low-cost selective surface
for sustained operation above 300°C so that the thermal emittance losses
would remain acceptable. However, releases of information have been difficult to interpret and compare as each company presents its information
differently.
The Novatec, Industrial Solar and SPG technologies all use an improved
version of the secondary reflector configuration first attempted by Paz, and
similar to that shown in Fig. 6.28. These three companies have recently
provided in the literature clearer information on heat loss and the efficiency
of conversion. In Bernhard et al., the heat loss from the SPG FRESDEMO
collector is given as the simple equation developed by the Fraunhofer
ISE in Freiburg; this was used by the author. A second closely matching
curve fit was also developed by the DLR but was not used here. The ISE
version is
q ′ = 0.011635 ⋅ ΔT 2
© Woodhead Publishing Limited, 2012
182
Concentrating solar power technology
6.28 Cross-sectional sketch of the Novatec receiver and secondary
reflector (courtesy Novatec Solar GmbH).
where q′ is the thermal energy lost in W/m2 of hot receiver surface, ΔT is
the difference in tenperature in °K between the output fluid temperature
and the ambient temperature near the collector. The ambient temperature
is not stated, but likely to be around 293–303°K. Bernhard et al. (2009)
present a graph (Fig. 3) of this heat loss compared to some actual data
points and also to a test of the Schott PTR-70 evacuated receiver. In the
graph, the evacuated receiver heat loss appears lower by a factor of 5 at
300°C above ambient. The same paper describes a measured optical efficiency of 0.62 but explains that the cover was not clean and makes a case
that a new cover would result in an optical efficiency of 65%, which is
accepted here as the SPG value.
Novatec give their heat loss as
Ploss = μ 0 ΔT + μ 1 ΔT 2
[6.1]
where Ploss is the heat loss in Watts per m2 of primary reflector, μ0 = 0.056 W/
(m2K) and μ1 = 0.000213 W/(m2K2). The ambient temperature is 313 K. ΔT
is the difference between ambient temperature and the output temperature
(normally 270°C). The optical efficiency is stated to be 67%.
Industrial Solar give their loss as simply the thermal loss per m2 of
primary reflector,
Ploss = 0.00043 W/(m 2 K 2 )
[6.2]
where the ambient temperature is taken as 303 K. The best optical efficiency is 66.3% at a sun angle 5° from the zenith – the central reflector
is shaded with the sun at zenith and the figure is consequently slightly
lower at solar noon. Industrial Solar has stated that the new design of
SPG has much lower losses but, as yet, there is no published data to verify
this.
© Woodhead Publishing Limited, 2012
Linear Fresnel reflector (LFR) technology
SP
AS
IS
N1
183
N2
Thermal loss
(W/m2 per m2 primary mirror)
160.0
140.0
120.0
100.0
80.0
60.0
40.0
20.0
0.0
250 260 270 280 290 300 310 320 330 340 350 360 370 380 390 400 410 420 430 440 450
Fluid output temperature (°C)
6.29 Comparison of estimated heat losses using information obtained
from several LFR manufacturers, present and future as described in
the text. The optical concentrations used were 54 for AS, 34 for SP, 25
for IS, and 52 for N1 and N2.
Figure 6.29 presents this heat loss data in a single graph with the ambient
temperature of 303 K used in each case. There are two Novatec-like cases;
the N1 using the published NOVA-1 heat loss and concentration data, and
the N2, using the same concentration but, like the SuperNOVA, a receiver
similar to a Schott-manufactured evacuated tube (Burkholder and Kutscher,
2009) with thick enough metal walls for the production of high pressure
superheated steam. The IS case heat loss is modelled on the published
thermal loss per m2 of the Industrial Solar system, which also uses a Schott
evacuated tube receiver, but uses about half the optical concentration of
the N2. Thus, the N2 heat loss modelled uses the Industrial Solar loss rate
multipled by 25/52 as a prediction of thermal loss, since the Industrial Solar
and Supernova systems have optical concentrations of 25 and approximately 52, respectively. The resulting N2 heat loss is about half the N1 case.
Therefore, the heat loss per m2 of reflector area of the IS system should be
quite similar to that of the N1 system, as is borne out in Fig. 6.29 up to
280°C. The published heat loss and relatively lower optical concentration
from SPG FRESDEMO are used in the SP case and heat loss is comparatively higher than N1 or IS, and much higher than N2.
Concerning Areva collection efficiency, Conlon (2011) describes the peak
thermal input from SSG4 to the heat transfer fluid as 7.86 MW at summer
solstice; the DNI assumed by Conlon is not stated but should lie in the range
of about ±5% of 950 W/m2. This would yield an optical efficiency of 7.86 MW/
(950 W/m2 × 11,261 m2/1,000,000) = 0.734 ± 0.0374. After thermal losses, the
mid-range efficiency is 7.30/10.7 = 0.682 ± 0.341. At 370°C and 104 bar pressure, the thermal loss is 0.56 MW, 49.7 W/m2 of reflector, and 5.2% of
© Woodhead Publishing Limited, 2012
184
Concentrating solar power technology
950 W/m2 DNI. Areva have not released the thermal loss per m2 of their
system over a range of temperatures, but an approximate heat loss curve
was constructed by the author that uses the (non-evacuated tube) N1 data
for convective thermal loss (adjusted for concentration ratio) and a radiative component from Industrial Solar that is adjusted for concentration
ratio togather with a higher assumed Areva emissivity that would produce
the known 49.7 W/m2 loss at an exit temperature of 370°C. The emissivity
adjustment factor that was necessary to produce this was 1.418, suggesting
a substantially higher emissivity for the Areva receiver than for the surface
in the Schott tubes, about 0.113 using this approximate method.
The AS loss curve in Fig. 6.29 shows this case, and suggests that the Areva
Solar heat loss might lie close to the Industrial Solar case, the increased
Areva concentration compensating for a higher surface emissivity and convective loss. The N2 heat loss, however, is about half that of AS, suggesting
that SuperNOVA heat loss might be significantly lower than for the Areva
SSG4. Emissivity largely dominates both heat loss systems at higher temperature, and the Novatec and Areva systems have similar optical concentration. The heat loss estimated for the Areva system seems plausible
but could be slightly higher or lower at other temperatures than 370°C
according to actual selective surface coating performance and convection
performance.
Does this heat loss advantage extend to efficiency? The Novatec efficiency is published but the Areva figures are less clear. Two graphs (Figs
6.30 and 6.31) have been shown in a recent Areva presentation as adjoining
slides (Venetos, 2010), with the latter also shown in Conlon (2011).2 Figure
6.31 was dated is September 2010, but both are almost certainly taken the
same day since a dropout in DNI occurs at 11.40 a.m. in both cases and at
no other time. From Fig. 6.30, the Areva line delivers a peak of around
Power (MW)
Measured vs. predicted performance
6
5
4
3
2
1
0
Measured
Predicted
11
12
13
14
15
16
Time of day
17
18
19
20
6.30 Measured thermal output from SSG4. This appears to be the
same day as indicated for the temperature and DNI in Fig. 6.31
(copyright AREVA Solar, used with permission).
2
Dr David Mills has not worked with AREVA Solar since June 2010 and has no direct knowledge of Areva solar’s present-day technology.
© Woodhead Publishing Limited, 2012
Linear Fresnel reflector (LFR) technology
185
Pressure [barg]
Bulk exit temperature [°C]
Saturation temperature [°C]
DNI
1000
900
800
700
600
500
400
300
200
100
0
20
DNI(W/m2)
Exit temperature (°C)
Data from SSG4: September 15, 2010
500
450
400
350
300
250
200
150
100
50
0
11
12
13
14
15
16
Time of day
17
18
19
Calculated peak thermal efficiency
6.31 Measured DNI and temperature from SSG4. This appears to be
the same day as indicated for the peak output in Fig. 6.30 because of
the 11.40 a.m. cloud event (copyright AREVA Solar, used with
permission).
AS η0 = 0.734
AS η0 = 0.75
N2
N1
IS
0.740
0.720
0.700
0.680
0.660
0.640
0.620
0.600
0.580
250 260 270 280 290 300 310 320 330 340 350 360 370 380 390 400 410 420 430 440 450
Fluid output temperature (°C)
6.32 Presentation of estimated peak thermal efficiency of AS-type
collectors with optical efficiency η0 = 0.75 and 0.734 (solid and dotted
lines), N2-type using η0 = 0.67 (large dashes and dots), IS-type
collectors with secondary reflectors with η0 = 0.663 (short dashes), and
N1-type with η0 = 0.67 (large dots). These curves are based on heat
losses as per Fig. 6.29.
5.5 MW(th) for about 90 minutes from a mirror area of 11,261 m2 at just
after 1.15 p.m., and in Fig. 6.32, the DNI averaged 950 W/m2 in the second
graph for about 90 minutes after 1 p.m.. There is a sizable time lag in such
long systems to allow fluid flow in such a long collector, hence the difference
in timings. The latitude of Bakersfield, the location of the Kimberlina array,
is 35.37°N, and using the US Naval Observatory website, the average cosine
© Woodhead Publishing Limited, 2012
186
Concentrating solar power technology
factor for the zenith angles averaged for the 90 minutes starting at 1p.m. on
15 September near Bakersfield was found to be about 0.7365 (USN, 2011).
Thus, the ‘averaged peak’ DNI of 950 W the incidant radiation around that
time was about 950 × 0.7365 = 700 W/m2 of collector area. For a 11,261 m2
array this amounts to 7.88 MW(th). The peak thermal efficiency of the unit
in September was thus calculated as 5.5/7.88 = 70% which should be
expected to be lower than the summer solstice figure already calculated,
but is in fact higher than the mid-range 68.2% solstice estimation based on
950 W/m2 of DNI. This suggests a higher optical efficiency, at least 5.5/7.88
+ 0.052 of heat loss = 0.750 and possibly higher. A higher figure than 0.75
is possible if the peak performance at solstice is based, for example, on
900 W/m2 DNI instead of 950.
Figure 6.32 shows peak performance modelling of systems with the
thermal loss data from Fig. 6.29 for the temperature range 250–450°C. In
the graph, two AS case curves are shown assuming optical efficiencies of
0.75 and the earlier estimation of 0.734. The results suggest that, in spite of
dramatically lower losses, there are some important downsides to evacuated
tube and secondary reflector use. In 2008, NREL tested the heat loss from
two PTR-70 tubes (Burkholder and Kutscher, 2009), which have a lower
heat loss than UVAC tubes (Burkholder and Kutscher, 2008). According to
Appendix IV of the NREL report (Burkholder and Kutscher, 2009), the
current 4 m PTR-70 evacuated tubes lose cumulatively 3.4% of optical
input due to light striking the end bellows. If we assume that the much
longer butt-welded non-evacuated tubes currently used in some designs like
Areva have a 0.4% connector optical loss (it could be even lower), then the
net optical deficit associated with evacuated tube use is assumed to be about
3%. This at first suggests a peak optical efficiency for N2 of close to 64%
instead of the N1 0.67. However, Industrial Solar already achieve η0 = 66.3%
using a similar Schott evacuated tube, so the author adjusted the assumption of optical efficiency of the calculated N2 system to 67% in Fig. 6.32
after discussion with a reviewer; the assumption is that compensating
improvements have been made, but the effect still limits optical efficiency
compared to an inverted AS1 or AS2 receiver. A second smaller disadvantage is that reflection losses may be greater for the cylindrical evacuated
tube cover tubes than for a flat cover; Morin et al. (2011) state that ‘rays
which are reflected at the secondary concentrator may pass several times
through the evacuated glass tube’. No SP case was calculated in Fig. 6.32
but if evacuated tube optical obstructions limit maximum optical efficiency,
then one might expect any revised future non-vacuum SP-type prototype
to do better by at least 3%, achieving about 70% optical efficiency. In
support, Morin et al. (2009; 2011) suggest that η0 = 0.705 is possible for SPG
type collectors with upgrading.
© Woodhead Publishing Limited, 2012
Linear Fresnel reflector (LFR) technology
187
The calculations are subject to several uncertainties, including the likely
optical cleanliness of the field, the relative blocking and shading of the
reflectors in different systems, and the relative absorptances of the tubes
used. Something favouring Areva in the calculation is that this is a peak
calculation with the sun essentially overhead; the Areva reflectors are larger
relative to the aperture with potentially greater off-angle astigmatism, but
this would not affect the zenith angle figure used. If it were a big problem,
then small Novatec style mirrors could be used, but Areva have not done
this. On the other hand, a scale drawing (fig. 2 in Conlon, 2011) shows a
widening gap between reflectors with greater distance from the receiver,
which would reduce shading and blocking at normal incidence and allow a
closer-packed field. It is obvious when seeing such a figure that keeping the
mirror field close to the receiver reduces image size and possible spillage;
Novatec and Industrial Solar use a constant gap that is larger between
reflectors closer to the receiver line, shifting the reflector field outward and
increasing average image size. This benefit for Areva may or may not persist
for off-angle of incidence solar radiation, but without comparative raytraces
and detailed measurements of reflector separations and tower heights, it
cannot be easily estimated in this review chapter.
Low concentration non-imaging LFR secondaries typically have an
average of about 0.5 reflections before reaching the receiver. For a clean
high-quality secondary reflector surface of 90–95% specular reflectance,
one might expect about between 5 and 2.5% loss respectively due to absorption or surface scattering in a reflector such as polished aluminium at 90%
or thin high temperature glass at 95%. However, primary reflector inaccuracy increases the number of reflections in the secondary, and we have
seen that evacuated tubes can cause a 3% extra optical loss due to obstruction, and there may be a small extra loss due to low angle incidence reflections from tubes. The results thus seem to suggest, but do not yet prove, that
the secondary losses in the reflector are a main source of the apparent
optical performance deficit, and obstructions by the evacuated tube have a
comparable negative effect. Note that all of the peak optical efficiencies
will drop for off-axis ray incidence, and that average daily optical efficiency
will be lower than the peak near noon on all days.
How do these estimates compare with trough collectors? In some respects,
a typical parabolic trough is a lot like an Industrial Solar system using
similar tubes and similar reflector area. A typical trough has an optical
efficiency of 0.75 according to NREL. Using Burkholder and Kutscher
(2009), the Schott PTR-70 evacuated receiver loses about 230 W/m length
at 400°C, a lower figure than its Solel UVAC3 tube competitor. Again
according to NREL, in a typical trough using the PT-70, there are 5.75 m2
of aperture per linear metre, so the loss is 40 W per m2 of primary field at
© Woodhead Publishing Limited, 2012
188
Concentrating solar power technology
400°C, 30% less than the A2 non-evacuated system but 40% higher than
the SuperNOVA. At 400°C, the peak trough thermal efficiency would be
around 0.71, slightly ahead of the author’s calculation of approximately
70% for the Areva LFR. Thus, in terms of peak thermal performance, the
parabolic trough and LFR are converging.
However, this is not the whole story. The conventional parabolic trough
arrangement is north-south tracking with plenty of room between the
reflectors to maintain the peak efficiency over much of the solar day. This
yields a higher capacity factor and annual output for the trough than is
possible for closely packed Fresnel designs, but also uses much more land.
The ‘capacity factor deficit’ is a big issue facing all LFR collectors and will
be discussed in the last section. On the other hand, LFR system cost is
lower.
6.8
Future trends
The companies who are currently the primary competitors in the LFR
market have a closely linked history, but have developed different priorities
and have chosen different technical development paths for their products.
There are a number of different issues to consider.
The original SHP (now Areva Solar) technology was initially developed
under the premises that
1. large high temperature evacuated tubes were expensive
2. small evacuated tubes were cheap but unsuitable for larger scale power
production for several reasons, including optical losses from groups of
receivers and coating limitations
3. hot cavity mirrors were unstable, air stable selective coatings were
unstable above 300°C but good performing surfaces near 300°C were
available
4. secondary reflectors can lose significant output in absorption
5. low temperature reheat turbines might be able to be used.
These days, highly efficient small reheat turbines have not come to the
market, hot mirror stability is improving, and secondary reflector performance and stability have been increased. However, the important efficiency
loss in LFR secondary reflectors remains and a peakier daily output than
troughs is a market problem. Evacuated tubes have the lowest loss rate,
they reduce heat stress on the secondary reflectors and also eliminate the
cost of a cover, but evacuated tubes are less optically efficient and may be
more prone to breakage than the simple cover solution. Another justification might be that evacuated tubes might be cheaper. That seems unlikely,
but Areva and SPG have not released details of their absorber coating
process or tube cost.
© Woodhead Publishing Limited, 2012
Linear Fresnel reflector (LFR) technology
189
But there is perhaps a broader question to ask. Do we need such high
temperature solutions everywhere in an LFR field? What is sometimes
overlooked is that in a solar collector system in which water is preheated,
boiled and superheated, the majority of the receiver system – perhaps 80%
– comprises the preheater and boiler, and here the peak temperatures for
these collector operations will lie in the temperature range below 370°C,
which may be accessible to low-cost selective absorber coating systems. In
other words, different receiver tube materials and coatings may be employed
for different parts of the solar array or even deployed together on one line.
If this comes to pass, then the low cost and high optical efficiency of present
day non-evacuated receivers may be applicable to most of the array, and
only a relatively small high temperature superheater section might need to
have specialist high temperature Schott-type evacuated tube solutions, if
they need them at all. SPG has announced that in large-scale implementation, their new high temperature glass reflector and absorber that will be
used in Mejillones should not be more expensive than the current midtemperature ones.
Both Areva and Novatec-Biosol have demonstrated automated mirror
production in the arrays commissioned in 2009 (Fig. 6.22). This presumably
is becoming standard. Installation is another matter. SHP/Ausra/Areva
have chosen a large reflector unit size, able to be installed by two staff using
a telehandler. In contrast, SPG and Novatec have chosen a small reflector
design able to be handled and installed by two or more staff. Each Ausra/
Areva reflector is about 36 m2 in area, about nine times that of Novatec
Solar and SPG reflectors. It is difficult to guess whether the 36 m2 mirror
approach of Areva Solar installed by telehandlers will win over the Novatec/
SPG approach of using many smaller mirrors. There are no doubt differences in the cost of shipping of different designs to the site, but this information is not available.
In terms of field configuration, the Areva arrays appear to be still growing
in size with each improvement, while the others have not changed significantly. This may be becasue the secondary reflector approach with evacuated tubes allows only one tube per receiver, restricting receiver size and
elevation. SPG has a much shorter line length (100 m) than Areva (384 m)
and Novatec (up to 986 m). This allows smaller scale arrays, but can lead
to increased optical end-effect losses. Longer arrays, however, increase
thermal expansion variations in the absorber tubes.
The notion of high temperature LFR operation becomes even more
powerful if one thinks about the combined impact of even higher LFR
optical concentration combined with better selective coatings. However,
there has been much confusion in the marketplace about concentration
ratios. Some trough manufacturers quote optical concentration as the ratio
between the mirror aperture and the absorber tube diameter, but this has
© Woodhead Publishing Limited, 2012
190
Concentrating solar power technology
no optical collection or thermal loss relevance, because no real or imaginary
surface along the diameter of the tube accepts or emits radiation. What
emits re-radiation is part, or all, of the absorber tube perimeter, depending
on the design. When viewed from the point of view of the standard definition of geometrical concentration, which might be defined as the ratio of
the collector aperture divided by the smallest surface area that can be
drawn around the re-radiation emitting portion of the receiver, today’s
LFRs at 50X+ are already substantially higher in concentration than the
25–30X typical of troughs.
Some studies have suggested that LFRs may not be efficient enough
to compete with troughs. However, these studies have been based around
the more public SPG system, not higher optical efficiency lower thermal
loss configurations like the Areva and Novatec systems. The origin of the
LFR performance deficiency relative to troughs is not so much in peak
performance; for example, Areva collectors are close to trough
peak performance at high temperature. It is rather the flatter output characteristic of north-south axis troughs during the day that increases their
plant capacity factor (CF). But in compensation, there seems to be acceptance that LFRs are considerably cheaper to build than PTCs. Two future
changes can profoundly affect this cost debate; field design and thermal
storage.
The current tendency to crowd the LFR reflectors together means that
each individual reflector does well at midday but less well at other times of
day when blocking occurs by adjacent reflectors. The recent paper by
Chavez and Collares-Periera (2010) suggests that optical concentration
close to 100 times is possible in ideal CLFR systems, and both CollaresPeriera and the author believe that practical systems of 75 times optical
concentration can be expected in the foreseeable future. For this, we need
wider systems. A wider spacing can yield an output per m2 which may be
superior if the receiver is designed to accept light from distant reflectors
without spillage. Current receivers are not designed to efficiently collect
from distant reflectors, but Chavez and Collares-Periera have discussed
such a system and the author has shown a design for a single tube (Mills,
1995c). Mirror-flipping, the defining characteristic of CLFRs discussed early
in this chapter, is also more effective if the receiver can collect more efficiently from distant field reflectors. The penalty of any spread-out system
will be some compromise of the ground area efficiency because the low
angles of radiation reflected to the tower demand wider spacing, but the
benefits of lower area cost and better distant mirror performance will likely
outweigh this. While the ‘spread-out’ system would incur some added secondary reflector absorption penalty for a higher concentration system, the
author believes this can be held to between 1 and 2% for future designs.
The lower relative receiver cost and better thermal performance of high
© Woodhead Publishing Limited, 2012
Linear Fresnel reflector (LFR) technology
191
concentration systems become very important at the higher operating temperatures of ∼550°C.
The other way to attack the capacity factor problem is to use storage.
Dersch et al. (2009) describe annual dumping fractions, the greatest of
which is caused in peak periods by a mismatch to turbine size. The peakier
LFR characteristic and higher dumping necessary when there is no storage
decrease the LFR plant capacity factor and total energy output per m2 relative to a trough. The price that north-south tracking axis troughs pay for
this advantage is a much larger land footprint, because the trough reflectors
need much more ground area than an LFR in order not to block and shade
one another as they track the sun’s elevation. Given the low price of arid
land, this is acceptable. However, the LFR capacity factor deficit suddenly
disappears with the advent of sufficient thermal storage, which removes
dumping and smooths output.
If we look briefly beyond LFRs, we see that Enel has recently built a salt
circulation trough plant using a new evacuated tube from the Archimede
company in Italy and Siemens (Archimede, 2011) to operate at 550°C, so
perhaps the worries of engineers can be resolved with respect to salt freezing in the field at night and overnight heat loss. If Archimede can produce
a successful salt storage receiver for trough plants, there is absolutely
nothing to stop an LFR using the same receiver technology, and possibly a
non-vacuum version can be designed also. Indeed, an LFR seems a superior
configuration to a linear salt plant, with fixed piping easily heated by reflectors and excellent drain-down possible. At temperatures of 550°C, heat loss
is important, and the higher concentration of LFRs would reduce that liability relative to a trough. If concentration can be extended to 75X in the
future as the author believes, then the LFR would require half the number
of receivers as a trough of the same annual output and have accordingly
lower heat loss. A company called Skyfuel (2011) currently is investigating
a molten salt LFR concept under a US DOE grant.
Molten nitrate salt is the current pre-eminent storage medium. It has
been demonstrated in troughs at lower temperatures and towers at 550°C
(Gould, 2011). At the higher temperatures, the amount of salt required to
store heat is substantially reduced through a larger temperature difference
(perhaps by a factor of 3) so that the CF difficulties previously alluded to
now largely disappear because the plant can follow the load precisely by
storing any excess energy in the molten salt tank. However, molten salt is
not the only high temperature circulation medium. The demonstrated
ability of LFR manufacturers to produce direct steam generation more
easily than trough collectors gives a powerful momentum for the development of 500–550°C superheated steam systems using the higher optical
concentration possible in an LFR system. If a storage system can be developed to store the output from a superheated steam LFR, then this may be
© Woodhead Publishing Limited, 2012
192
Concentrating solar power technology
the lowest cost system, and easier to manage than salt-filled linear collectors. The DLR is already working on such systems.
If significant thermal storage becomes standard, then it would seem that
the collector design that collects the most energy for the least cost will be
the commercial winner, regardless of the collector thermal output profile.
Current LFRs are strong on the low-cost aspect but weaker on the annual
energy per collector area side. However, in a short time they have begun
to approach trough performance and if capacity factor issues are reduced
through storage or field design, they may prevail.
6.9
Conclusions
The LFR market is already highly competitive, and all of the major LFR
manufacturers are now progressing with versions that surpass most current
trough output temperatures (about 400°C), and later will approach current
tower technology temperatures. The recent superheating results above
450°C by SPG, Areva and Novatec herald this future. In this chapter, comparisons have sometimes been difficult to make without access to what
continues to be proprietary commercial information, but it is clear that
development of both major collector types is rapid and each can use any
storage technology developed for troughs. No decision on the correct design
for an LFR can yet be made, although peak performance is approaching
that of troughs in the case of the non-secondary reflector LFR. What can
perhaps be said with greater confidence is that LFR energy production cost
is dropping rapidly compared to trough technology.
Towers on their own promise high temperature through very high optical
concentration, but low field and O&M costs seem to be elusive. In the
recent competitive shortlisting for the Australia flagship project, LFRs and
troughs proceeded to the short list but no tower did. Troughs offer high field
efficiency, proven technology, and a flat daily output pattern, but downsides
are low concentration and high field costs. With LFRs and CLFRs, the
advantages of high temperature, fixed pipes for high pressure operation,
low losses using selective coatings and medium concentration, low field cost,
and relatively compact land usage are potentially available in one package.
Francia’s inspiration 50 years ago was a profound one. Yet many significant improvements are still possible in what is a young, exciting technology.
This coming decade will see many new LFR developments take their place
in the market.
6.10
References
Appleyard D (2010), ‘Chilling in the heat of the Doha sun’, Renewable Energy
World, December.
© Woodhead Publishing Limited, 2012
Linear Fresnel reflector (LFR) technology
193
Archimede (2011), Solar Receiver Tubes for Thermodynamic Power Plants, Archimede Solar website, http://www.archimedesolarenergy.com/en_home.asp
(accessed 15 January 2012).
Areva (2012), CLFR performance and specifications, at http://www.areva.com/EN/
operations-3641/compact-linear-fresnel-reflector-technology.html (accessed 6
January 2012).
Ausra (2007), CLFR Coal saver plants. Statement by Ausra, the predecessor of
Areva Solar, at http://www.solarpaces.org/Tasks/Task1/clfr_coal_saver.htm
(accesssed 15 January 2012).
Bernhard R, Laabs HJ, de Lalaing J, Eck M, Eickhoff M, Pottler K, Morin G,
Heimsath A, Georg A, Häberle A (2008), ‘Linear Fresnel collector demonstration
on the PSA, Part I – Design, construction and quality control’, Proc. 14th International Symposium on Concentrated Solar Power and Chemical Energy Technologies, 3–7 March, Las Vegas.
Bernhard R, Lalaing J, Kistner R, Eck M, Eickhoff M, Feldhoff JF, Heimsath A,
Hülsey H, Morin G (2009), ‘Linear Fresnel collector demonstration at the PSA
– Operation and investigation’, Proc. 15th International Symposium on Concentrated Solar Power and Chemical Energy Technologies, 14–18 Sept., Berlin.
Buffon G-L (1830), ‘Burning instruments’, in The Edinburgh Encyclopaedia, D.
Brewster (ed.). 135–137.
Burkholder F and Kutscher C (2008), Heat Loss Testing of Solel’s UVAC3 Parabolic
Trough Receiver. National Renewable Energy Laboratory Report NREL/TP-55042394, January.
Burkholder F and Kutscher C (2009), Heat Loss Testing of Schott’s 2008 PTR70
Parabolic Trough Receiver. National Renewable Energy Laboratory Report
NREL/TP-550-45633, May.
Chavez J and Collares-Periera M (2010), ‘Etendue-matched two-stage concentrators with multiple receivers’, Solar Energy, 84 (2), 196–207.
Conlon W (2011), ‘Superheated steam from CLFR solar steam generators’, Proc.
16th International Symposium on Concentrated Solar Power and Chemical Energy
Technologies, 21–24 Sept., Perpignan, France.
de Lalaing J. (2009), Interview with Jaques de lalaing by Beyond Zero Emissions,
August. Available at http://w.w.w.beyondzeroemission.org/media/radio/countjaques-de-lalaing-solar-power-groups-linear-fresnel-solar-thermal-090817.
Dersch J, Morin G, Eck M, Häberle A (2009), ‘Comparison of linear Fresnel and
parabolic trough collector systems – system analysis to determine break even
costs of linear fresnel collectors’, Proc. 15th International Symposium on Concentrated Solar Power and Chemical Energy Technologies, 14–18 Sept., Berlin.
Dey CJ, Mills DR, Morrison GL (2000), ‘Operation of a CLFR research apparatus’,
Proc. New Zealand Solar Energy Society Annual Conference – From Fossils to
Photons, Brisbane.
Di Canio DG, Tretyl WG, Jur FJ, Watson CD (1979), ‘Line-focus solar thermal
central receiver research study’, Final Report 1977–79 DOE/ET/20426-1, OSTI
ID: 5535434, FMC Corporation, Santa Clara, CA.
Eck M, Uhlig R, Mertins M, Häberle A, Lerchenmüller H (2007), ‘Thermal load of
direct steam-generating absorber tubes with large diameter in horizontal linear
Fresnel collectors’, Heat Transfer Engineering, 28 (1), 42–48.
Feuermann D (1993), ‘Analysis and evaluation of the solar thermal system at the
Ben-Gurion Sede Boqer Test Center for solar electricity generating technologies’,
© Woodhead Publishing Limited, 2012
194
Concentrating solar power technology
Final Report, Contract No. 88169101, Israel Ministry of Energy and Infrastructure, Jerusalem, July.
Feuermann D and Gordon JM (1991), ‘Analysis of a two-stage linear Fresnel reflector solar concentrator’, Transactions of the ASME, 113, 272–279.
Francia G (1968), ‘Pilot plants of solar steam generation systems’, J. Solar Energy,
12, 51–64.
Gould W (2011), ‘Solar Reserve’s 565 MWt molten salt power towers’, Proc. 16th
International Symposium on Concentrated Solar Power and Chemical Energy
Technologies, 21–24 Sept., Perpignan, France.
Häberle A, Mertins M, Lerchenmüller H, Heinzel V (2004), ‘Geometry optimization
of Fresnel collectors with economic assessment’, Proc. 14th International
Sonnenforum – Proceedings 1, EuroSun2004, 1-918–1-925.
Industrial Solar (2011a), ‘Mirroxx is now Industrial Solar’. Pressbox Industrial Solar
GmbH, 4 May, at Hhttp://www.pressebox.com/pressreleases/industrial-solargmbh/boxid/416345 (accessed 31 May 2011).
Industrial Solar (2011b), ‘Technical Data; Industrial Solar linear Fresnel collector
LF-11’, 30 March, at http://www.industrial-solar.de/cms/uploads/media/IS_technical_data_300311.pdf (accessed 31 May 2011).
Mertins M, Lerchenmüller H, Häberle A (2004), ‘Geometry optimization of Fresnel
collectors with economic assessment’, Proc. 14th International Sonnenforum,
EuroSun2004, 20–24 June, Freiburg, Germany.
Mills DR (1995a), ‘Proposed solar cogeneration powerplant for 2000 Olympics’,
Proc. Solar ’95 Renewable Energy: the Future is Now, 33rd Annual Meeting,
465–473, International Solar Energy Society Congress (also presented at the
Harare ISES meeting, but no proceedings were forthcoming).
Mills DR (1995b), Australian patent #694335, plus filings in USA, Europe, China.
Mills DR (1995c), ‘Two-stage collectors approaching maximal concentration’, Solar
Energy, 54 (1), 41–47.
Mills DR and Dey CJ (1999), ‘Transition strategies for solar thermal power generation’, Proc. International Solar Energy Society Congress, Jerusalem, 5–9 July.
Mills DR and Morrison GL (1999), ‘Compact linear Fresnel reflector solar thermal
powerplants’, Solar Energy, 68 (3), 263–283.
Mills DR, Morrison G, Le Lièvre P (2004a), ‘Design of a 240 MWe solar thermal
power plant’, Proc. 14th International Sonnenforum, EuroSun2004, 20–24 June,
Freiburg, Germany.
Mills DR, Morrison G, Le Lièvre P (2004b), ‘Lower temperature approach for very
large solar power plants’, Proc. 12th SolarPACES International Symposium,
Oaxaca, Mexico, October.
Mills DR, Morrison G, Pye JD, Le Lièvre P (2006), ‘Multi-tower line focus Fresnel
array’, Journal of Solar Energy Engineering, 128 (1), 118–120.
Morin G, Platzer W, Eck M, Uhlig R, Häberle A, Berger M, Zarza E (2006), ‘Road
map towards the demonstration of a linear Fresnel collector using a single tube
receiver’, Proc. of SolarPACES 13th International Symposium on Concentrated
Solar Power and Chemical Energy Technologies, 20–23 June, Seville, Spain.
Morin G, Dersch J, Eck M, Häberle A, Platzer W (2009), ‘Comparison of linear
Fresnel and parabolic trough collector systems – influence of linear Fresnel
collector design variations on break even cost’, Proc. 15th International Symposium on Concentrated Solar Power and Chemical Energy Technologies, 14–18
Sept., Berlin.
© Woodhead Publishing Limited, 2012
Linear Fresnel reflector (LFR) technology
195
Morin G, Dersch J, Häberle A, Platzer W, Eck M (2011), ‘Comparison of linear
fresnel and parabolic trough power plants’, Solar Energy, doi:10.1016/j.
solener.2011.06.020, in press.
NOVA-1 (2011a), ‘The future. Today. PE 1 – World’s first Fresnel Solar Power Plant
in commercial operation’, at http://www.novatecsolar.com/files/mne0911_PE1_
broschure_english.pdf (accessed 15 January 2012).
NOVA-1 (2011b), Technical data sheet, at http://www.novatecsolar.com/files/
mntd1102_technical_data_sheet.pdf http://www.novatecsolar.com/files/mn1102_
nova1_brochure_eng_web.pdf (accessed 15 January 2012).
Paul C, Bachelier C, Teichrew O, Weckert C (2011), ‘Power plant concepts between
CSP solar-only developments and integrated solar combined cycles based on the
latest commercially available Fresnel collectors’, Proc. 16th International Symposium on Concentrated Solar Power and Chemical Energy Technologies, 21–24
Sept., Perpignan, France.
PEM (2011), FNEG Solar, Power Engineering Magazine, 30 March, at http://www.
power-eng.com/articles/2011/03/areva-fneg-solar.html (accessed 15 January
2012).
Pye JD, Morrison GL, Behnia M, Mills DR (2003), ‘Modelling of the cavity heat
transfer for the compact linear Fresnel reflector’, Proc. Australia and New Zealand
Solar Energy Society Conference, Adelaide.
Reynolds D, Jance M, Behnia M, Morrison GL (2001) ‘An experimental and
computational study of the heat loss characteristics of a trapezoidal cavity
absorber’, Proc. of International Solar Energy Society Solar World Congress,
Adelaide.
Shaner WW and Duff WS (1979), ‘Solar thermal electric power systems: comparison
of three line focus collectors’, Solar Energy, 22 (1), 49–61.
Skyfuel (2011), High temperature linear Fresnel, at http://www.skyfuel.com/index_
main.html#/OUR%20TECHNOLOGY/HIGH%20TEMPERATURE%20
LINEAR%20FRESNELR/ (accessed 15 January 2012).
Solar Dawn (2011), Website regarding the 250 MW flagship solar project in
Queensland, at http://solardawn.com.au/ (accessed 15 January 2012).
SPG (2008), Continuous steam production at 450°C, 15 September, at http://www.
solarpowergroup.com/169.0.html (accessed 15 January 2012).
SPG (2011), Continuous steam production at 450°C. Solar Power Group
web address http://www.solarpowergroup.com/113.0.html (accessed 15 January
2012).
SuperNOVA (2011), ‘Fresnel specialist Novatec Biosol turns to superheated steam
to boost efficiency by 50%’, Renewable Energy Magazine, 4 June, at http://www.
energias-renovables.com/energias/renovables/index/pag/solar_thermal_electric/
botid/73/colright/solar_thermal_electric/tip/articulo/pagid/12273/title/
Fresnel%20specialist%20Novatec%20Biosol%20turns%20to%20super
heated%20steam%20to%20boost%20efficiency%20by%2050porciento/len/en
(accessed 15 January 2012).
USN (2011), Sun or Moon Altitude/Azimuth Table, US Naval Oceanography Portal,
at http://aa.usno.navy.mil/data/docs/AltAz.php (accessed 1 June 2011).
USOTA (1978), Application of solar technology to today’s energy needs, Chapter
VIII, Collectors, 270. Report by the Congress of the United States Office of Technology Assessment (USOTA). Available for free download at http://www.princeton.edu/∼ota/disk3/1978/7802/780212.pdf.
© Woodhead Publishing Limited, 2012
196
Concentrating solar power technology
Venetos M (2010), ‘Compact Linear Fresnel Reflector Technology for Power Augmentation’, Proc. Solar Power International, Los Angeles.
Winston R and Rabl A (1976), ‘Ideal concentrators for finite sources and restricted
exit angles’, Applied Optics, 15 (11), 2889–2883.
Zhang Q-C and Mills DR (1992), ‘New cermet film structures with much
improved selectivity for solar thermal applications’, Appl. Phys. Lett., 60 (5),
545–547.
© Woodhead Publishing Limited, 2012
7
Parabolic-trough concentrating solar power
(CSP) systems
E. Z A R Z A M OYA, CIEMAT –
Plataforma Solar de Almería, Spain
Abstract: This chapter gives an overview of the parabolic-trough
collector (PTC) technology, the technology most widely used in solar
thermal power plants today. It includes a brief history of the earliest
parabolic-troughs and a description of the first commercial projects
implemented in the 1980s, the main parameters and basic equations of a
typical PTC, design criteria, operation and maintenance issues and
expected technology improvements in working fluids and solar collectors
that could be implemented in the short to medium term.
Key words: solar energy, parabolic-trough collectors, solar concentration,
linear solar collectors, solar thermal power plants.
7.1
Introduction
A parabolic-trough collector (PTC) is a linear-focus solar collector, basically composed of a parabolic-trough-shaped concentrator that reflects
direct solar radiation onto a receiver or absorber tube located in the focal
line of the parabola (see Fig. 7.1). The larger collector aperture area concentrates reflected direct solar radiation onto the smaller outer surface
of the receiver tube, heating the fluid that circulates through it. The solar
radiation is thus transformed into thermal energy in the form of sensible
or latent heat of the fluid. This thermal energy can then be used to feed
either industrial processes demanding thermal energy (e.g., food industry,
petro-chemical industry, etc.) or Rankine cycles to produce electricity with
a steam turbine in a ‘solar thermal’ power plant.
With today’s technology, parabolic troughs can deliver useful thermal
energy up to 398°C. The main limitation on the maximum temperature is
imposed by the thermal oil currently used as the working fluid, because it
quickly degrades above 398°C. However, research in new fluids promises
higher temperatures close to 500°C in the mid term.
7.1.1
Historical development
The first document describing the optical properties of a parabolic-trough
collector was written by the Greek mathematician Diocles in the second
197
© Woodhead Publishing Limited, 2012
198
Concentrating solar power technology
Steel structure
Parabolic trough reflector
Absorber tube
7.1 A typical parabolic trough collector.
century BC. Diocles explained that a parabolic mirror reflects the solar rays
towards a common point located at a specific place (i.e., the focal point of
the parabola). However, more than 20 centuries had to pass before the first
solar system using parabolic-trough collectors was actually erected. The first
graphically documented parabolic-trough solar collector was designed and
built by Swedish engineer, John Ericsson, at the end of the nineteenth
century. This collector had a solar radiation collecting surface of 3.25 m2
and produced saturated steam for a small 373 W steam engine. In 1883, the
original design using reflectors made of polished sheet metal was improved
by Ericsson with the installation of flat strips of silvered-glass reflectors in
a frame providing the parabolic shape with a 3.35 × 4.9 m aperture. It produced saturated steam at 0.24 MPa to drive a 120 rpm steam engine with a
single 0.153 m diameter 0.2 m stroke piston. Any possibility of commercial
development died with Ericsson in 1889 because he always kept the technical details secret.
Later, in 1913, the American engineer, Frank Shuman, designed and built
a parabolic-trough collector solar plant in Egypt that produced 0.1 MPa
saturated steam to drive a steam engine for pumping irrigation water for a
farming community in Meadi, near Cairo. Shuman had the financial support
of British investors, and with the valuable technical advice of British professor, C.V. Boys, improved the overall plant design. Figure 7.2 shows one of
the five collectors installed at Meadi in 1913. Every collector was 62.17 m
long with a 4.1 m wide parabola, for a total solar field collecting surface of
1274.5 m2, and 40% collector efficiency. Water was pumped at a flow rate
of 380 l/s with the energy supplied by the solar field. The reflectors and
receiver tubes were supported by a metal structure on a concrete foundation. The photograph in Fig. 7.2 was taken from the book by Hans Gunther
(1922). Although the quality is not very good due to the age of the original
© Woodhead Publishing Limited, 2012
Parabolic-trough concentrating solar power (CSP) systems
199
7.2 One of the five parabolic-trough collectors installed by Frank
Shuman at Meadi (Egypt) in 1913.
photograph and the digitalization required for reprinting here, the similarity between the parabolic-trough collectors designed by Shuman in 1912
and the collectors installed in solar thermal power plants today is clear.
However, the First World War and the entry of oil into the energy market
resulted in no more such plants being built at that time.
Twenty years later, in 1935, C.G. Abbot converted solar energy into
mechanical power using a parabolic-trough collector (PTC) and a 0.37 kW
steam engine (Pytlinski, 1978). After that, there was no further outstanding
PTC development in the twentieth century until a renewed interest in solar
energy following the 1973 oil price crisis when three PTC prototypes were
developed and tested in the USA by Sandia National Laboratories, Honeywell International Inc. and Westinghouse, and a detailed cost study was
carried out in 1979 (Shaner and Duff, 1979). A short time later, in the 1980s,
several other companies developed new PTC designs and entered the
market with small industrial process heat applications and small solar
thermal power plants. Table 7.1 gives the details of four small demonstration solar thermal power plants built in the USA, Japan, Spain and Australia
at that time.
Among others, the Acurex Solar Corp. (PTC models Acurex 3001 and
Acurex 3011), Suntec Systems Corp./Excel Corp. (PTC models IV and 360),
Solar Kinetics Corp. (PTC models T-700 and T-800), General Electric, Honeywell Inc. and Jacobs Del. Corp., manufactured and marketed a number
of PTCs during the early 1980s (Fernandez-García et al., 2010). Simultaneously, the only two-axis PTC that has ever been marketed, the Helioman
3/32, was developed by the German company Maschinenfabrik AugsburgNürnberg (MAN).
© Woodhead Publishing Limited, 2012
200
Concentrating solar power technology
Table 7.1 Details of demonstration trough-based solar thermal power plants
built during the early 1980s
Net electric power
Total aperture area
Heat transfer fluid
Effective storage
capacity
Duration of service
Coolidge
(USA)
Sunshine
(Japan)
IEA-DCS
(Spain)
STEP-100
(Australia)
0.15 MWe
2,140 m2
Synthetic oil
5 MWth
1 MWe
12,856 m2
Water/steam
3 MWe
0.5 MWe
7,622 m2
Synthetic oil
0.8 MWe
0.1 MWe
920 m2
Synthetic oil
117 MWth
1980–1982
1981–1984
1981–1987
1982–1985
Although other companies, among them the American company Industrial Solar Technologies (IST), entered the market with successful PTC
designs later on, they were mainly for industrial process heat applications.
The most outstanding event related to PTC technology in the twentieth
century was the design and implementation of the nine SEGS (Solar Electricity Generating System) plants in the Mohave Desert (California, USA)
by LUZ International Limited from 1984 (SEGS-I) to 1990 (SEGS-IX).
Table 7.2 gives the specifications for the nine SEGS plants, while Fig. 7.3
shows one of the SEGS plants and Fig. 7.4 illustrates the operating principle
schematically. A synthetic heat transfer oil is pumped through the trough
array and heated by concentrated solar radiation as it circulates through
the receiver pipes. This oil is then used to produce steam in heat exchangers
before being circulated back to the solar field. The steam is used in a conventional steam turbine-based electricity generating plant. Although some
hot oil-based energy storage was provided in the first plant, the SEGS
systems overall rely on natural gas firing to provide continuous operation
when the sun is not available.
With a total electrical output of 354 MW and more than two million
square meters of parabolic-trough collectors, the SEGS plants have been
an invaluable aid in the improvement and commercial deployment that
parabolic-trough collectors are now experiencing at the beginning of the
twenty-first century (Harats and Kearney, 1989). The fact that these plants
are still in daily operation after so many years provides a high level of
technical credibility and confidence for both investors and promoters. But
the SEGS plants have not only contributed to the commercial take-off of
parabolic-troughs by acting as a ‘technology showcase’ for new projects
promoted in the twenty-first century, they have also made possible the
existence of an industrial capability to at least partly satisfy the initial
demand for essential components, such as receiver tubes and reflectors.
© Woodhead Publishing Limited, 2012
© Woodhead Publishing Limited, 2012
Net electric power
Efficiency in solar
mode (%)
PTC model
Total aperture area
(m2)
Solar field inlet/outlet
temperatures (°C)
Synthetic oil (HTF)
Duration of service
SEGS 2
1985–
present
30 MWe
29.40
LS1/LS2
188,990
248/320
M-VP1
SEGS 1
1984–
present
13.8 MWe
28.5
LS1
83,000
241/307
ESSO500
M-VP1
248/349
LS2
230,300
1986–
present
30 MWe
30.60
SEGS 3
M-VP1
248/349
LS2
230,300
1986–
present
30 MWe
30.60
SEGS 4
M-VP1
248/349
LS2
233,120
1987–
present
30 MWe
30.60
SEGS 5
M-VP1
293/393
LS2
188,000
1988–
present
30 MWe
37.70
SEGS 6
Table 7.2 Specifications for the nine SEGS plants built by the company LUZ International
M-VP1
293/393
LS3
194,280
1988–
present
30 MWe
37.50
SEGS 7
M-VP1
293/390
LS3
464,340
1989–
present
80 MWe
37.60
SEGS 8
M-VP1
293/390
LS3
483,960
1990–
present
80 MWe
37.60
SEGS 9
202
Concentrating solar power technology
7.3 View of one of the SEGS plants.
395 °C Oil
Superheated steam (104 bar/380 °C)
Steam turbine
G
Solar field
Steam
generator
Gas-fired
oil heater
Condenser
Water
Degasifier
Reheated steam (17 bar/371 °C)
295 °C Oil
Water Preheater
Reheater
Oil expansion vessel
7.4 Schematic configuration of a typical SEGS plant.
The SEGS plants were developed because of the favorable conditions
defined by the legal framework implemented in the United States as a
consequence of the steep rise in oil prices in the 1970s. However, these
government incentives for renewable systems were cut and became insufficient when the oil price fell again in the 1980s, thus making installation of
more SEGS plants unfeasible (Lotker, 1991).
Since early in this century, government incentives have again been set up
in some countries (e.g., tax credits in the USA, a favorable feed-in tariff in
Spain and Algeria, etc.), promoting a multitude of solar thermal power
plant projects. Most of these new projects are based on parabolic-trough
collectors, because long-term successful track records of the SEGS plants
establish them as the ‘least risk’ and most readily financeable solution. This
© Woodhead Publishing Limited, 2012
Parabolic-trough concentrating solar power (CSP) systems
203
surge of new solar power plants has led to a significant investment in R&D,
and the development of many new parabolic-trough collector designs and
new factories for key components (e.g., reflectors and receiver tubes) in the
USA and Spain.
7.2
Commercially available parabolic-trough
collectors (PTCs)
7.2.1
Large PTCs
One of the achievements of LUZ International was the development of
three reliable, durable PTC designs, called the LS1, LS2 and LS3, which
were successfully implemented and operated in the SEGS plants. Although
the LS1 installed by LUZ at the SEGS-I plant in 1984 had a unit length of
50.2 m and a parabola width of 2.5 m, which was similar in size to other
designs developed during the early 1980s, it soon became evident that
bigger PTCs would have to be developed for larger CSP plants. This was
the reason why LUZ developed the LS2 and LS3 designs, with aperture
areas of 235 and 545 m2 per collector, respectively. After the demise of LUZ,
one of the barriers to the installation of large solar fields with parabolic
troughs was lack of a suitable PTC. In view of this, a European consortium
composed of industry, engineering firms and R&D centers was formed in
1998 to develop a new PTC suitable for large CSP plants. The result was
the Eurotrough-100 (ET-100) and EuroTrough-150 (ET-150) designs, which
were then improved, leading to their successor, the SKAL-ET, the collector
installed at the ANDASOL-I plant in Spain in 2007. Figure 7.5 shows the
7.5 Steel structure of the Eurotrough-100 collector design.
© Woodhead Publishing Limited, 2012
204
Concentrating solar power technology
Table 7.3 Parameters of the Eurotrough-150 collector design
Parabola width (m)
Overall length of a single collector (m)
Length of every module (m)
Number of parabolic trough modules
Outer diameter of the steel receiver pipe (m)
Inner diameter of the steel receiver pipe (m)
Collector aperture area (m2)
Mirrors reflectivity
Steel receiver pipe absorptance
Intercept factor
Receiver pipe glass cover transmittance
Peak optical efficiency
5.76
147.5
12.27
12
0.07
0.065
822.5
0.93
0.95
0.90
0.95
0.75
steel structure of the ET-100 design, and Table 7.3 gives the parameters of
the ET-150 collector design.
Highly precise assembly of the steel structural profiles is required to
achieve perfect parabolic shape of the concentrator and overall structural
rigidity, while at the same time keeping assembly cost low. Thus, although
not all PTC designs require assembly jigs, most of the modern designs do
require them to meet the design tolerances, which are usually around
±1 mm. Figure 7.6 shows the final check of an assembly jig for the
EuroTrough collector design.
The collector design shown in Fig. 7.5 has a central space frame, called
torque box, which provides good rigidity and prevents torsion, which is very
important to ensure good PTC optical and geometrical performance under
moderate wind speeds (below 35 km/h). Figure 7.6 also shows how a
Eurotrough parabolic-trough module looks when all its components have
been mounted in the assembly jig. In large parabolic-trough solar fields, two
or three assembly lines are used in parallel to shorten the construction time.
Another PTC design approach is the replacement of the torque box by a
central steel tube (called the torque tube). However, assembly of the steel
mirror support frames on this central tube must also be highly accurate.
Figure 7.7 (bottom) shows a detail of the collector developed by the Spanish
company, Albiasa Solar (www.albiasasolar.com), using the torque tube
concept. The collectors developed by the Spanish companies, SENER
(www.sener.es) and URSSA (www.urssa.es), in 2006–2010 also have a
torque tube instead of a torque box.
The main advantage of PTC torque box designs is their usually better
performance and rigidity under wind loads, while their main disadvantage
is their higher assembly cost due to the number of steel profiles which
require high-precision assembly. Torque tube designs, on the other hand,
are usually somewhat cheaper, but are subject to more deformation from
© Woodhead Publishing Limited, 2012
Parabolic-trough concentrating solar power (CSP) systems
205
7.6 Check out of the assembly jig for EuroTrough collectors (top) and
a parabolic-trough module assembled in the jig (bottom).
gravity (bending) and wind loads (torsion) than those using a torque box.
In any case, rigidity in wind loads of both PTC designs is good enough to
keep deformation within reasonable limits.
There are also commercial PTC designs which provide a good stiff structure without a torque box or a torque tube, which are replaced by a metallic
space frame, such as the one with a 430.8 m2 collector aperture area used
by Solargenix and Acciona in the Nevada Solar One Plant, or the American
Skyfuel company’s SkyTrough collector, which has a collector aperture area
of 690 m2 (www.skyfuel.com). The design used by Solargenix and Acciona
is also notable in that it is one of the few PTC designs that does not use an
accurate jig for assembly, but instead relies on the accuracy of the preformed parts and fasteners to generate an accurate shape. Figure 7.7 (top)
shows the space frame developed by the company Gossamer Space Frames
for Solargenix and successfully used in the CSP plant Nevada Solar One,
and which provides good torsional rigidity and performance in wind loads.
© Woodhead Publishing Limited, 2012
206
Concentrating solar power technology
7.7 PTC designs using a space frame (top) and torque tube (bottom)
to provide good torsional rigidity.
Other means of providing a stiff structure in order to reduce costs and
simultaneously maintain good rigidity and assembly accuracy are under
study. However, these innovative designs were only in the prototype stage
at the end of 2010.
Since the size and number of components in the structure of large PTC
collectors used in solar thermal power plants make correction of errors
almost impossible after the structures have been installed in the solar field,
an effective quality control procedure during the whole assembly process
is extremely important to ensure that design tolerances and performance
are met. In addition to a highly accurate assembly of the steel structures,
the alignment of the parallel rows of collectors in the solar field must also
be accurate to avoid solar tracking errors, especially when open-loop tracking systems based on mathematical calculation of sun coordinates are used.
© Woodhead Publishing Limited, 2012
Parabolic-trough concentrating solar power (CSP) systems
207
Any misalignment would introduce a tracking error and a corresponding
reduction in the concentrated solar radiation finally reaching the receiver
pipes.
Although the main type of reflector used in commercial PTC designs is
the curved back-silvered low-iron thick-glass mirror, it can also be made of
polished sheet metal, or silver or aluminum-coated films that can be laminated onto a rigid parabola-shaped substrate. There are several suppliers of
this type of polymer film reflector. VikuitiTM marketed by 3M (http://www.
solutions.3m.com) and ReflecTech® marketed by Reflectech Inc. (http://
www.reflectechsolar.com) are two examples of polymer film reflectors for
solar applications. However, experimental data about their outdoor performance and durability are still limited and not as plentiful as for glass mirrors.
7.2.2
Small PTCs
Although most of the R&D effort in the late twentieth century and early
in the twenty-first century has been devoted to big parabolic-trough collectors for large solar thermal power plants, new smaller troughs have also
been developed for process heat applications with temperatures below
300°C. These collectors can provide process heat to a wide range of applications replacing natural gas, such as crop drying and food preparation. Industrial processes such as biofuel production, water purification, desalination
and absorption-chiller air conditioning for commercial and industrial buildings are within the capability of these small parabolic-trough collectors.
IST was very active in developing several collectors and installing a
number of commercial plants in the USA. In 2007 IST was purchased by
the Spanish industrial group, Abengoa, and their IST-PT1 and IST-RMT
collectors were then marketed by Abengoa Solar IST (www.abengoasolar.
com). Both collectors have non-evacuated receiver pipes with a blackchrome coating and glass envelope, and maximum working temperatures
of 288°C and 204°C, respectively.
Another small PTC for process heat applications is the PTC-1800 collector developed by the German-Turkish company, SOLITEM (www.solitem.
de). The maximum working temperature of this collector is 200°C and the
collector aperture area is 36.65 m2.
The Australian company, NEP Solar Pty Ltd (www.nep-solar.com), also
developed a small PTC, the PolyTrough-1200, suitable for temperatures up
to 230°C. The 24 m long, 1.2 m wide PolyTrough-1200 collector consists of
composite reflector panels with aluminum sheets which can be mounted
either on roofs or at ground level.
SOPOGY (www.sopogy.com), with head offices in Honolulu (Hawaii),
has developed a 3.7 m long, 1.35 m wide parabolic-trough collector module
marketed as Soponova 4.0, which is suitable for both process heat
© Woodhead Publishing Limited, 2012
208
Concentrating solar power technology
7.8 The Soponova 4.0 parabolic-trough concentrator developed by
SOPOGY.
applications and electricity generation. Figure 7.8 shows a Soponova 4.0
collector module. In 2010, eight solar power plants had been built around
the world by SOPOGY. The most recent was in Keahole (Hawaii), inaugurated in December 2009, with a 2 MWe rated power output. SOPOGY
developed a small solar thermal power plant concept scalable from 250 kWe
to over 20 MWe based on Soponova 4.0 trough-collector arrays. This concept
of a scalable solar thermal power plant was patented by SOPOGY under
the trade name MicroCSPTM.
Table 7.4 gives the technical parameters of the small parabolic-trough
collectors marketed by Abengoa Solar IST, SOLITEM and SOPOGY.
7.2.3
Receivers
The typical PTC receiver tube is in fact composed of two concentric pipes,
an inner steel pipe containing the working fluid and an outer glass tube
surrounding the steel pipe. The glass tube is made of low-iron borosilicate
glass to increase its transmittance for solar radiation. The outer surface of
the steel pipe has an optically selective surface with a high solar absorptance and low emittance for thermally generated infra-red radiation. The
principles of such surfaces are discussed in detail in Chapter 15. The glass
tube is usually provided with an anti-reflective coating to achieve a higher
solar transmittance and better annual performance.
Receivers for parabolic-trough collectors can be classified as either evacuated or non-evacuated. Evacuated receivers are commonly used for temperatures above 300°C because they have a high vacuum (i.e., 10−5 mbar)
between the steel pipe and the glass cover, thus reducing thermal losses and
© Woodhead Publishing Limited, 2012
Parabolic-trough concentrating solar power (CSP) systems
209
Table 7.4 Technical parameters of Abengoa IST, SOLITEM and SOPOGY
parabolic troughs
IST-RMT
IST-PT1
PCT-1800
Soponova 4.0
Structure
Steel/
aluminium
Steel/
aluminium
Aluminium
Steel/
aluminium
Reflectors
Polished
aluminium
Aluminium
acrylic
Al-coating
0.5 mm
Aluminium
Trough
module size
3.66 × 1.13 m
6.1 × 2.9 m
5.09 × 1.8 m
3.66 × 1.52 m
Receiver
tube
Nonevacuated,
with blackchrome
coating
Nonevacuated,
with
blackchrome
coating
Nonevacuated,
with selective
coating
Nonevacuated,
with selective
coating
Maximum
temperature
204°C
288°C
∼200°C
260°C
increasing the overall efficiency of the PTC, especially at higher operating
temperatures. Figure 7.9 shows a typical evacuated receiver. The glass cover
of these receivers is connected to the steel pipe by means of stainless steel
expansion bellows which not only compensate for the different thermal
expansion of glass and steel when the receiver tube is working at nominal
temperature, but also provide a tight annular gap between both tubes to
make the vacuum. One end of these expansion bellows is directly welded
to the outer surface of the steel pipe, while the other end is connected to
the end of the glass cover by means of a glass-to-metal welding. Shown in
Fig. 7.9 are chemical ‘getters’ placed in the gap between the steel receiver
pipe and the glass cover to absorb gas molecules passing from the fluid to
the annulus through the steel pipe wall. Since the evacuated receivers are
expensive (about 850 c/unit in 2010) due to their technical complexity, they
are used only for higher temperatures, when good thermal efficiency is
required and the high cost is compensated by a higher thermal output.
At the end of 2010, there were only three manufacturers of evacuated
PTC receivers: Schott, Siemens and ASE. Most of the parabolic-trough
solar thermal power plants implemented around the world until 2009 had
receivers manufactured by either the Israeli company, Solel (purchased in
2009 by Siemens, www.energy.siemens.com), or the German company,
Schott (www.schottsolar.com). In 2009, a third manufacturer, the Italian
company, Archimede Solar Energy (ASE, www.archimedesolaenergy.com),
announced that they were launching a new receiver tube called HEMS08,
© Woodhead Publishing Limited, 2012
210
Concentrating solar power technology
Glass pin to evacuate the air Vacuum between the glass cover
and the steel pipe
Glass-to-metal welding
Steel pipe with
‘Getters’ to keep and maintain
selective coating
the vacuum
Expansion bellows
Glass cover
Siemens’s design
Schott’s design
ASE’s design
7.9 A typical evacuated receiver for parabolic-trough collectors.
suitable for fluids up to 550°C. The first plant using HEMS08 receivers was
the Archimede Plant, located in Syracuse (Italy) and ready to operate in
2010 using molten salt (a mixture of sodium and potassium nitrate) as the
receiver working fluid.
Figure 7.9 shows how these three manufacturers join the glass cover and
the inner steel pipe by means of flexible bellows. The glass-to-metal welding
used to connect the glass cover to the flexible bellows is a weak point in the
receiver tube and has to be protected from the concentrated solar radiation
to avoid high thermal and mechanical stress that could cause the welding to
crack. An aluminum shield is therefore usually placed over the flexible
bellows to protect the welding. Table 7.5 shows the technical parameters of
the receivers manufactured by the Schott, Siemens and ASE companies.
Non-evacuated receivers are suitable for applications with a working
temperature below 300°C, because thermal losses are not so critical at these
temperatures. Although non-evacuated receivers are also composed of an
inner steel pipe and a glass cover, they have neither vacuum between the
steel pipe and its glass cover nor glass-to-metal welds. Selective coatings
used for non-evacuated receivers are simpler than those used for evacuated
receivers. Black-chrome or black nickel coatings are commonly used
because they are cheap and easy to produce.
Due to manufacturing constraints, maximum receiver tube length is
usually less than 5 m, so they are connected in series up to the total length
of the PTC. Evacuated receivers are usually welded, while non-evacuated
receivers are usually connected by special threaded joints.
© Woodhead Publishing Limited, 2012
Parabolic-trough concentrating solar power (CSP) systems
211
Table 7.5 Technical parameters of the receivers commercialized by Schott,
Siemens and ASE
Schott PTR-70
Siemens
UVAC-2010
ASE HEMS08
Solar absorptance
Solar transmittance
Thermal emittance
≥0.95
≥ 0.96
≤0.1 at 400°C
≥0.96
≥ 0.96
≤0.09 at 400°C
Steel pipe inner/
outer diameters
Thermal losses
Glass cover
Active length ratio
at 350°C
Maximum fluid
temperature
70/65 mm
stainless steel
250 W/m at 400°C
Borosilicate
>96%
70/65 mm
stainless steel
n.a.
Borosilicate
96.4%
≥0.95
n.a.
≤0.1 at 400°C
≤0.14 at 580°C
70/65 mm
stainless steel
230 W/m at 400°C
Borosilicate
n.a.
400°C
400°C
550°C
7.3
Existing parabolic-trough collector (PTC) solar
thermal power plants
Figure 7.10 shows the configuration of a typical parabolic-trough solar
thermal power plant with thermal oil (the HTF1, or heat transfer fluid technology). Such plants, which were the industry standard at the end of 2010,
can be divided into three subsystems:
1. The solar field. This is where the direct solar radiation is collected and
converted into thermal energy in the form of the sensible heat of the
working fluid between the solar field and the power block or thermal
storage system. The solar field is composed of parallel rows of solar
collectors.
2. The thermal energy storage system. Shown at the centre of Fig. 7.10, this
subsystem is not essential to plant operation. However, a thermal storage
system provides clear benefits to plant operation, because it not only
increases annual hours of operation, and thereby, the amount of electricity produced, but also improves plant dispatchability and enhances plant
operation during cloud transients by sending the entire thermal energy
produced by the solar field to the storage system instead of to the power
block. This operating strategy avoids damage to the steam turbine from
unstable steam parameters by acting as a thermal buffer between the
1
The term heat transfer fluid is used because the fluid transfers heat from the solar field to
the point of use and typically (for power generation) is different from the power cycle working
fluid.
© Woodhead Publishing Limited, 2012
212
Concentrating solar power technology
Thermal energy storage system
395 °C Oil
Superheated steam (104 bar/380 °C)
Steam turbine
(385 °C)
Molten salts
Power block
G
(hot tank)
Solar field
Steam
generator
Condenser
Deaerator
(295 °C)
Molten salts
(cold tank)
Reheated steam 17 bar/371 °C
295 °C Oil
Preheater
Reheater
Oil expansion vessel
7.10 Configuration of a typical solar thermal power plant with
parabolic-trough collectors.
solar field and the power block, and its thermal inertia filters any
temperature-related instability in the solar field outlet. However, even
without thermal storage, the significant thermal inertia provided by the
solar field piping and the steam generator of a large solar thermal power
plant can provide enough thermal energy to run the power block at
steady output for a few minutes after a cloud affects the solar field. The
thermal storage system shown in Figure 7.10 is a two-tank (cold and
hot) molten-salt system and its operation is explained in Section 7.6.
3. The power block. This is where the thermal energy delivered by either
the solar field or the storage system is converted into electricity by
means of a steam Rankine cycle. Since the power block for a parabolictrough system is similar to conventional power plants (e.g., water pumps,
wet cooling systems, steam turbine, electricity generator, de-aerator and
water/steam heat exchangers), the required maintenance work is also
the same or very similar.
Although not shown in Fig. 7.10, these solar power plants may use an
auxiliary gas-fired oil heater to allow plant operation when solar radiation
is not available and there is no thermal energy available in the storage
system. This heater is usually installed in parallel to the solar field because
the experience gained with the SEGS plants showed that plant operation
is more difficult if the gas-fired heater is installed in the power block and
produces the superheated steam for the steam turbine directly. Although
overall plant efficiency would be higher with this configuration because it
avoids thermal losses in the oil circuit, the change from solar to fossil plant
operating mode would be much more complicated.
© Woodhead Publishing Limited, 2012
Parabolic-trough concentrating solar power (CSP) systems
213
On a typical clear day, the solar field tracking is started when direct solar
radiation is in the 100–300 W/m2 range. During the first few minutes, the
thermal oil is recirculated through the solar field until it reaches the nominal
outlet temperature, and then sent to either the thermal storage system or
to the steam generator to start the power block. During daylight hours in
summer, the solar field of plants with a thermal storage system delivers
enough energy to keep the power block running at full load and at the same
time charge the storage system. Since the solar collectors are usually
installed with their rotation axis oriented north-south, the thermal output
of the solar field on clear winter days is much less than in summer, and
thermal storage is not used often because the thermal output of the solar
field on clear winter days is only enough to feed the power block. This is
the reason why the use of the thermal storage system in winter is usually
limited to partly cloudy days.
The profitability achieved in a few countries due to public incentives in
the form of feed-in tariffs or tax credits implemented during the first decade
of the twenty-first century promoted a multitude of solar thermal power
plant projects, most of them with parabolic-trough collectors, because the
successful track record of the SEGS plants established it as the most financeable, ‘least-risk’ technology. Although the world financial crisis of 2008 was
a serious barrier to developing all the projects that were initially proposed
at that time, many of them became a reality and a significant number of
CSP plants with parabolic troughs were in operation at the end of 2010.
Table 7.6 gives the list of CSP plants in operation with parabolic troughs at
the end of 2010, producing a total electrical output of 876 MWe. Since, due
to the strong industrial competition, most of the owners of these plants are
rather reluctant to publish annual performance data, such information
could not be included in Table 7.6.
7.4
Design of parabolic-trough concentrating solar
power (CSP) systems
This section explains the most important parameters of a typical parabolictrough collector (PTC) and makes recommendations to achieve a good
performance. It also describes the PTC energy balance and solar field
design criteria.
7.4.1
Basic PTC parameters
Commercial PTC designs for solar thermal power plants are 100 m to 150 m
long, and have a parabola width of about 6 m, which provides an aperture
area of 550 m2 to 825 m2 (approx.). Larger PTC designs are under development and could become commercially available in the mid term.
© Woodhead Publishing Limited, 2012
© Woodhead Publishing Limited, 2012
Granada, Spain
Nevada, USA
Puertollano,
Spain
Badajoz, Spain
Badajoz, Spain
Badajoz, Spain
Siracusa, Sicily,
Italy
Arizona, USA
Hawaii, USA
Badajoz, Spain
Caceres, Spain
Badajoz, Spain
Laghoutat,
Algeria
Kuraymat, Egypt
100 (2 × 50)
64
50
50
50
50
5
2
50
50
50
20
40
Andasol I and II
Nevada Solar One
Ibersol Puertollano
Alvarado I
Extresol 1
La Florida
Archimede solar power
plant
Saguaro Solar Power
Station
Keahole Solar Power Plant
La Risca
Planta Termoel. de Majadas
La Dehesa
Hassi’R Mel-1
Kuraymat
1
Seville, Spain
150 (3 × 50)
California, USA
354
Solar Energy Generating
Systems
(SEGS-I, IX)
Solnova-I, III and IV
Location
Nominal
Power (MWe)
Plant name
NREA
Arizona Public
Services
Keahole Solar Power
ACCIONA
ACCIONA
SAMCA
ABENER + NEAL
Acciona Solar Power
ACS
SAMCA
ENEA
Acciona Solar Power
Iberdrola
ACS
Abengoa
Sunray Energy, (SEGS
I + II)
FPL Energy (SEGS
III–IX)
Owner
Table 7.6 List of CSP plants with parabolic troughs in operation at the end of 2010
Inaugurated in December 2009
Completed in July 2009
Completed in November 2010
Completed in July 2010
150 MWe ISCCS plant with 20 MWe
solar. Completed in Summer 2010
470 MWe ISCCS plant with 40 MWe
solar. Completed in December 2010
Completed July 2009
Completed in February 2010
Completed in July 2010
ISCC with molten-salt heat storage
Completed in July 2010
2006
Solnova I (May 2010)
Solnova III (May 2010)
Solnova IV (August 2010 )
Andasol I (2008)
Andasol II (2009)
Completed in June 2007
Completed in May 2009
Collection of nine units ranging from
14 to 80 MWe and
operating since the 1980s
Remarks
Parabolic-trough concentrating solar power (CSP) systems
215
Vector perpendicular
to the aperture plane
Sun vector
Incidence angle, ϕ
Parabolic-trough concentrator
Collector aperture plane
7.11 Correct positioning of a parabolic-trough concentrator.
As solar concentrating devices, parabolic-trough collectors require solar
tracking systems to modify their position with the changing apparent sun
position in the sky from sunrise to sunset. Movement of this type of solar
collector has only one degree of freedom, on-axis rotation. The concentrator must always reflect and concentrate the beam solar radiation onto the
receiver tube, and a proper concentration is not possible if the rotation
angle is not right. Figure 7.11 shows how direct solar radiation has to reach
the collector aperture plane in order to be properly reflected onto the
receiver tube. The position of the PTC must be such that the sun vector, the
collector focal line and the vector perpendicular to the collector aperture
plane are on the same plane. The angle defined by the two vectors shown
in Fig. 7.11 is called the incidence angle. It strongly affects the amount of
incident solar flux available on the PTC aperture plane, as the smaller the
incidence angle, the more incident solar flux can be reflected and converted
into useful thermal energy in the receiver pipe. Since diffuse solar radiation
falling on the Earth’s surface has no specific direction, this component of
the solar radiation is useless to a PTC, because it cannot be reflected onto
the receiver tube by the concentrator. The fundamentals of CSP systems
have been described in general terms in Chapter 2, so some of this material
is repeated here with additional specific details for PTC systems.2
The most important PTC parameters are the geometric concentration
ratio, acceptance angle, rim angle and peak optical efficiency. These parameters are explained in the following paragraphs.
The geometric concentration ratio, Cg, is the ratio between the collector
aperture area and the total absorber tube area (see Fig. 7.12). This concentration ratio is usually about 25, although theoretically, the maximum is on
2
Note that the symbols used for variables are similar but not identical to those in Chapter 2.
© Woodhead Publishing Limited, 2012
216
Concentrating solar power technology
β
Aperture area Ac = l · la
W
Receiver tube
y
b/2
Steel receiver tube outer
diameter, d0
l
Cg =
Wa · l
p · d0 · l
(a)
(b)
7.12 (a) Geometric concentration ratio, Cg and (b) acceptance angle,
β and aperture angle, ψ of a parabolic-trough collector.
the order of 70. High concentration ratios are associated with higher
working temperatures. The Geometric concentration ratio, Cg, is given by
Eq. [7.1]:
Cg =
W ⋅l
W
=
π ⋅ do ⋅ l π ⋅ do
[7.1]
where do is the outer diameter of the receiver steel pipe, l is collector length,
and W is the parabola width.
The acceptance angle, β, is the maximum angle that can be formed by
two rays on a plane transversal to the collector aperture in such a way
that, when they are reflected by the parabolic mirrors, they intercept the
absorber pipe. The wider the collector acceptance angle is, the less accurate
the sun tracking system has to be, as the collector will not need to update
its position as frequently. Small acceptance angles are associated with high
concentration ratios, which require the installation of very accurate solartracking systems and, consequently, higher costs. The minimum acceptance
angle is 32′ (0.53°), which is the average solid angle with which the solar
disk is seen from the Earth. Therefore, any PTC with an acceptance angle
smaller than 32′ would always lose a fraction of the direct solar radiation.
In fact, recommended acceptance angles for commercial PTCs are in the
range 1–2°. Smaller angles would demand very accurate and more expensive solar-tracking systems, while wider angles would lead to small concentration ratios and, therefore, lower working temperatures. So most
commercial PTC designs have acceptance angles within the range 1–2°,
with geometric concentration ratios of 20 to 30.
© Woodhead Publishing Limited, 2012
Parabolic-trough concentrating solar power (CSP) systems
217
Beam solar radiation
γ
Absorber glass cover
(with transmissivity t)
Steel absorber pipe
(with absorptivity a)
Parabolic reflector
(with reflectivity r)
7.13 Optical losses in a parabolic-trough collector.
The rim angle, ψ, which is directly related to the concentrator arc length,
can be calculated from Eq. [7.2] as a function of the parabola focal distance,
f, and aperture width, W:
8 ⋅ f ⋅W
= tan ψ
W 2 − 16 ⋅ f 2
[7.2]
Usual rim angles are in the range 70–110°. Smaller rim angles are not advisable because they reduce the aperture surface. Rim angles over 110° are
not cost-effective because they require the whole reflecting surface to be
enlarged without significantly increasing the area of the aperture plane.
Optical losses are very important in parabolic-trough collectors because
they are about 25% of the total solar flux incident on the PTC aperture
plane. Optical losses are associated with the following four parameters (see
Fig. 7.13):
•
•
Reflectivity, ρ, of the collector reflective surface. Since the reflectivity of
the parabolic-trough concentrator is less than 1, only a fraction of the
incident solar flux is reflected towards the receiver tube. Typical reflectivity values of clean silvered glass mirrors are around 0.93.
Intercept factor, ϒ . A fraction of the direct solar radiation reflected by
the mirrors does not reach the active surface of the receiver pipe due
to either microscopic imperfections of the reflectors, macroscopic errors
in the parabolic-trough concentrator shape (e.g., inaccuracies during
assembly), mechanical deformation of the PTC, flexible bellows, or
© Woodhead Publishing Limited, 2012
218
•
•
Concentrating solar power technology
shadowing by the receiver tube supports. If microscopic imperfections
are neglected, the intercept factor with a 0° incidence angle of the beam
solar radiation can be considered the result of three parameters: geometrical errors in the parabolic-trough concentrator shape, γ1, shadowing by the flexible bellows, γ2, and mechanical deformation (i.e., bending
+ torsion) of the support structure, γ3. All of the above parameters cause
either some rays to be reflected at the wrong angle or block some of the
reflected rays, preventing them from intercepting the steel absorber
tube. All these losses are globally quantified by the intercept factor, ϒ .
This optical parameter is typically within the 0.91–0.93 range for highquality PTCs because γ1 ≅ 0.97, γ2 ≅ 0.96 and γ3 ≅ 0.99.
Transmissivity of the glass cover, τ. The steel receiver tube is inserted in
a glass cover to reduce thermal losses. A fraction of the direct solar
radiation reflected by the mirrors onto the glass cover of the receiver
pipe is unable to penetrate it. The ratio of the radiation passing through
the glass cover to the total incident radiation on it, is the transmissivity,
τ. of the glass. It is typically τ = 0.93, and can be increased up to 0.96 by
anti-reflective coatings applied on both sides of the glass cover.
Absorptivity of the receiver selective coating, α. This parameter quantifies
the amount of energy absorbed by the steel receiver pipe over the total
radiation reaching its outer wall. This parameter is typically 0.95 for
receiver pipes with a cermet selective coating, and slightly lower for
pipes coated with black nickel or chrome.
Multiplication of these four parameters (reflectivity, intercept factor,
glass transmissivity, and absorptivity of the steel pipe) when the incidence
angle of the solar flux onto the PTC aperture plane is 0° gives the peak
optical efficiency of the PTC, ηopt,0°:
ηopt,0° = ρ × γ × τ × α ϕ = 0°
[7.3]
ηopt,0° is usually in the range 0.74–0.79 for clean, good-quality parabolictrough collectors.
Taking the four optical parameters included in the peak optical efficiency
into consideration, it clearly represents the percentage of the beam solar
radiation reaching the PTC aperture plane finally absorbed by the receiver
pipe when the incidence angle is 0°.
The incidence angle of the beam solar radiation, ϕ, affects the four optical
parameters mentioned above and the useful aperture area of the collector.
This effect is quantified by the incidence angle modifier, K(ϕ), which includes
all optical and geometric losses in a PTC due to an incidence angle greater
than 0°. So the percentage of the beam solar radiation reaching the PTC
aperture plane with the incidence angle ϕ that is finally absorbed by the
receiver pipe is the result of multiplying the peak optical efficiency, ηopt,0°,
by the incidence angle modifier, K(ϕ).
© Woodhead Publishing Limited, 2012
Parabolic-trough concentrating solar power (CSP) systems
ηopt,ϕ ≠ 0° = ηopt,0° K (j )
219
[7.4]
The incidence angle modifier depends directly on the incidence angle and
is usually given by a polynomial equation such that it is equal to 0 for high
ϕ (≥85°), and to 1 for ϕ = 0°. Thus the incidence angle modifier for an LS-3
PTC, for instance, is given by:
K(ϕ ) = 1 − 2.23073E-4 × j − 1.1E-4 × j 2
+ 3.18596 E-6 × j 3 − 4.85509 E-8 × j 4
K (ϕ ) = 0
(0° < ϕ < 80°)
(85° < ϕ < 90°) [7.5]
Thermal losses are also very important in parabolic-trough collectors
because they signifi
. cantly affect the overall collector efficiency. Total thermal
losses in a PTC,. Qcollector_ambient, are due to radiative heat losses from the steel
receiver tube, Qabsorber_ambient
, and convective and conductive heat losses from
.
it to its glass cover, Qabsorber_glass. Though these heat losses are governed by
the well-known mechanisms of radiation, conduction and convection, they
are experimentally calculated for every PTC and receiver pipe design by
operating the collector under real solar conditions at several temperatures.
These experimental results are then processed to find the thermal loss equation as a function of the steel receiver tube temperature and the ambient
air temperature. Sometimes, thermal losses are calculated as a function of
the working fluid temperature, the ambient air temperature and the solar
flux incident onto the aperture plane. In any case, the result is a mathematical equation delivered by the PTC manufacturer to. the solar field designer
for calculating overall thermal losses in the PTC, Qcollector_ambient.
Today’s high temperature PTCs are provided with evacuated receiver
pipes, thus avoiding convection losses between the steel pipe and its glass
cover.
7.4.2
Energy balance in a PTC
Figure 7.14 illustrates the energy balance in a typical
. PTC. The solar energy
flux incident on the aperture plane of the PTC, Qsun_collector, is. shown on the
left, and the useful thermal energy delivered by the PTC, Qcollector_fluid, is on
the right. The three sources of energy loss in the PTC explained in Section
7.4.1 and shown at the bottom of Fig. 7.14 are:
•
•
•
optical losses due to mirror reflectivity, intercept factor, glass transmissivity and solar absorptance of the receiver tube when the solar radiation incidence angle is equal to 0°, ηopt,0°
additional optical and geometrical losses due to an incidence angle
greater than 0°, K(ϕ); these additional losses do not exist when the
incident angle is equal to 0° because K(ϕ = 0°) = 1
.
thermal losses from the receiver pipe to the ambient, Qcollector_ambient.
© Woodhead Publishing Limited, 2012
220
Concentrating solar power technology
hsystem
hopt, 0°
hth
K(j)
QQ, collector→fluid
QQ, sun→collector
hopt, 0°
K(j)(j>0°)
QQ, collector→ambient
7.14 Energy balance in a parabolic-trough collector.
Taking into consideration the energy balance illustrated in Fig. 7.14, overall
PTC efficiency, ηsystem, is calculated .as the ratio of the net thermal output
power delivered by the collector, Qcollector_fl
uid, to the solar energy flux inci.
dent on the collector aperture plane, Qsun_collector, according to Eqs [7.6], [7.7]
and [7.8]:
ηsystem =
QQ,collector → fluid
QQ, sun→ collector
[7.6]
Q sun _ collector = Ac ⋅ Ed ⋅ cos (ϕ )
[7.7]
⋅ ( hout − hin )
Q collector _ fluid = m
[7.8]
where Ac is the collector aperture surface, Ed is the direct solar irradiance,
.
ϕ is the incidence angle, m is the fluid mass flow through the collector
receiver tube, hin is the fluid specific mass enthalpy at the collector inlet,
and hout is the fluid specific mass enthalpy at the collector outlet.
Equation [7.8] is used when the PTC is in operation and mass flow and
temperatures are known. Since the fluid mass flow and the inlet and outlet
temperatures are not known during the solar field design phase, the expected
net thermal output has to be theoretically calculated from the energy
balance shown in Fig. 7.14, and direct solar irradiance, ambient air temperature, incidence angle and PTC optical, thermal and geometrical parameters
using Eq. [7.9]:
Q collector _ fluid = Ac ⋅ Ed ⋅ cos (ϕ ) ⋅ ηopt,0° ⋅ K (ϕ ) ⋅ Fe − Q collector _ ambient
[7.9]
All the parameters used in Eq. [7.9] were explained above, except for the
soiling factor, Fe, which is calculated as the ratio between average PTC
mirror reflectivity during real operation and the nominal reflectivity when
the PTC is completely clean. So, for instance, if the nominal mirror reflectivity of a PTC is 0.93 and the PTC is washed when reflectivity falls
to 0.89, the average mirror reflectivity is (0.93 + 0.89)/2 = 0.91 and Fe =
© Woodhead Publishing Limited, 2012
Parabolic-trough concentrating solar power (CSP) systems
221
0.91/0.93 = 0.978. Fe for commercial PTC solar fields is usually in the
0.95 < Fe < 1 range.
The collector aperture area, optical peak efficiency, incident angle modifier, and thermal losses versus the PTC working conditions and ambient air
temperature in Eq. [7.9] are supplied by the PTC manufacturer, while the
beam solar irradiance, soiling factor and incident angle are defined by the
solar field designer taking local weather conditions, site latitude and longitude and the solar field mirror washing procedure to be used by the plant
owner into consideration.
The optical and thermal quality of modern parabolic-trough collectors
used in solar thermal power plants is very high. Evacuated receiver pipes
significantly reduce thermal loss to less than 35 kW for an average fluid
temperature of 375°C in a complete 150 m long and 828 m2 aperture collector. The useful thermal output (Eq. [7.9]) for a beam solar radiation of
925 W/m2 and an incidence angle of 15° is about 450 kW at an ambient
temperature of 25°C.
The high optical quality of reflectors used in high-temperature PTCs,
along with the high accuracy ensured by the assembly procedure, lead to
an excellent peak optical efficiency, ηopt,0°, in the 0.74–0.79 range. This high
optical efficiency and the low thermal loss ensured by the evacuated receivers achieve a high overall efficiency (Eq. [7.6]) of about 70% for working
temperatures of 375°C. Since most of the high-temperature PTCs in solar
thermal power plants employ similar solar reflectors and evacuated receivers and have similar assembly tolerances, their performance is also very
similar. Smaller PTC designs for industrial process heat applications are less
efficient, because they do not have evacuated receivers and the optical
quality of the solar reflectors commonly used (polished-aluminum metallic
reflectors) leads to lower overall efficiencies, usually in the 0.5–0.65 range
at a working temperature of 250°C. However, performance can vary a lot
from one collector model to another.
7.4.3
Design of parabolic-trough solar fields for CSP plants
A typical parabolic-trough collector field (Fig. 7.15) is composed of parallel
rows of collectors. Each row, in turn, is composed of several collectors connected in series so that the working fluid circulating through the receiver
pipe is heated as it passes from the inlet to the outlet of each row.
There are three stages in parabolic-trough collector solar field design:
•
•
Stage 1: Define the design point, which is the set of assumed design
values.
Stage 2: Calculate the number of parabolic-trough collectors to be connected in series in each parallel row.
© Woodhead Publishing Limited, 2012
222
Concentrating solar power technology
7.15 Parallel rows in a solar field with parabolic-trough collectors.
•
Stage 3: Calculate the number of parallel rows to be installed in the solar
field.
For Stage 1 (definition of the design point) the solar field designer must
consider not only local weather conditions, but also the parameters of the
PTC design chosen for the plant and any customer specifications. Since
solar radiation is a non-constant energy source, it is evident that the thermal
output delivered by a solar field will not be constant either. This means that
the thermal power delivered by the solar field will sometimes be higher and
sometimes lower than the design one. The solar field produces the nominal
thermal output when working conditions and parameters are the same as
assumed for the design point. Parameters and working conditions to be
defined for the design point are:
•
•
•
•
•
•
collector orientation
design point date (month and day) and time
site location (latitude and longitude)
direct solar irradiance and ambient air temperature for the selected date
and time
total thermal output power to be delivered by the solar field
soiling factor of the solar field
© Woodhead Publishing Limited, 2012
Parabolic-trough concentrating solar power (CSP) systems
2
4
6
8
10
12
14
16
18
Pcollector→fluid (KW)
300
300
250
250
200
200
North-south orientation
East-west orientation
150
150
100
100
50
50
0
0
2
4
6
8
10
12
14
16
18
0
20 22 24
350
0
20 22 24
Solar time
(a) Power output on a Summer clear day
2
4
6
8
350
Pcollector→fluid (KW)
0
350
223
10 12 14 16 18 20 22 24
350
300
North-south
orientation
300
250
East-west
orientation
250
200
200
150
150
100
100
50
50
0
0
2
4
6
8
0
10 12 14 16 18 20 22 24
Solar time
(b) Power output on a Winter clear day
7.16 Daily thermal output of a EuroTrough-100 parabolic-trough
collector located in Southern Spain, on a Summer (a) and Winter (b)
clear day, and with two different orientations of its axis (north-south
and east-west).
•
•
•
solar field inlet/outlet temperatures
solar collector working fluid
nominal fluid flow rate.
The collector orientation is vey important because the seasonal performance of a PTC depends strongly on this parameter. Figure 7.16 shows the
daily thermal output of a typical PTC located in Southern Spain, on a clear
day in summer (Fig. 7.16 (a)) and winter (Fig. 7.16 (b)), and with two different axis orientations (north-south and east-west). The PTC design
assumed in the figure is a Eurotrough-100. It may be clearly observed that
with a north-south orientation there are significant seasonal variations
between summer and winter, so the peak thermal output in winter is less
than 50% of the peak thermal output in summer, while a PTC with eastwest orientation has similar peak thermal output in winter and summer.
Figure 7.16 also shows that the thermal energy delivered by the north-south
collector in winter is about 50% of the thermal energy delivered by the
east-west collector, while the opposite is true in summer. However, since
there are more clear days and hours of daylight in summer, the north-south
orientation has a higher yearly thermal output. Although the seasonal differences in collectors oriented North-South decrease as the site nears the
Equator, Fig. 7.16 is valid for most countries where solar thermal power
plants are being installed, because they are in latitudes within the range
20–40°N.
Taking into consideration the seasonal performance associated with
different orientations, the solar field designer must select the best orientation for the solar plant, depending on the thermal output demand and the
site’s geographic coordinates. Since current solar thermal power plants need
to maximize their yearly electricity production, they all use north-south
© Woodhead Publishing Limited, 2012
224
Concentrating solar power technology
oriented collectors, because this allows the CSP plant to provide more
electricity at peak hours in the summer months, when the electricity demand
increases due to air-conditioners, and electricity is more expensive, thus
increasing revenues from electricity sales. However, in principle, any intermediate angle between north-south and east-west is also possible.
The design point date and time usually chosen is a summer day at noon
(for instance 21 June in the northern hemisphere) because the thermal
output of north-south oriented collectors is maximum at that date and time.
If the solar field were designed for a winter’s day, a huge thermal storage
system would be necessary to avoid dumping thermal energy in summer
when the solar field delivers more than 200% of the mid-winter power.
Once the design point date and time has been chosen by the designer,
and the geographic location has been defined by the client, the designer
must assume design point weather conditions (e.g., direct solar radiation
and ambient air temperature) usual at that site, date and time. Weather
stations located nearby or a synthetic meteorological year obtained from
satellite data can be used for this purpose.
The rated thermal output power to be delivered by the solar field and its
inlet/outlet temperatures are imposed by the thermal industrial process or
power block to be fed. For solar thermal power plants, the nominal inlet/
outlet temperatures are 293°C/393°C (approx.) because thermal oils currently used as the working fluid rapidly degrade above 398°C, and overall
plant efficiency is maximized with a temperature step of about 100°C in the
solar field. For industrial process heat applications, the solar field outlet
temperature needs to be at least 15°C higher than the steam temperature
demanded by the process to be fed. So if the industrial process requires
steam at 300°C, the oil temperature at the solar field outlet must be about
315°C. This difference is necessary to compensate thermal losses between
the solar field outlet and the steam generator inlet, and provides a temperature differential and compensates the boiler pinch point, which is on the
order of 5–10°C.
The selection of the working heat transfer fluid (HTF) for a PTC solar
field is also important in the design phase. A single-phase liquid provides
the best heat transfer coefficients and stable operation. Thermal oil is commonly used in parabolic-trough collectors for temperatures above 200°C.
Water maintained as a liquid by pressurizing beyond the saturation pressure
requires high pressure inside the receiver tubes and piping at these operating temperatures, requiring stronger joints and piping, and thus raising the
price of the collectors and solar field. For temperatures below 200°C, either
a mixture of water/ethylene glycol or pressurized liquid water can be used
as the working fluid because only a moderate pressure is required to keep
the fluid in liquid phase. Direct conversion of liquid water into highpressure saturated or superheated steam in the receiver pipes of the solar
© Woodhead Publishing Limited, 2012
Parabolic-trough concentrating solar power (CSP) systems
225
collectors, the so-called direct steam generation process, has been studied
thoroughly at the Plataforma Solar de Almería (PSA) since 1998 and its
technical feasibility has been proven, so its marketing will soon be a reality
(Zarza et al., 2004).
Several thermal oils are suitable for use as HTFs for parabolic-trough
collectors. One of the key parameters to be considered when choosing the
appropriate type of oil is the maximum bulk temperature at which the
manufacturer guarantees oil stability. The oil most widely used in parabolictrough collectors for temperatures up to 395°C is a eutectic mixture of
73.5% diphenyl oxide/26.5% diphenyl (Dawtherm A or VP-1 thermal oil).
The main problem with this type of oil is its high crystallization temperature
(12°C), which requires installation of an auxiliary heating system if oil
piping temperature could drop below this temperature limit. Since the
boiling temperature at 1013 mbar is 257°C, the oil circuit must be pressurized with nitrogen, argon or another inert gas when oil is heated above this
temperature. Blanketing of complete oil circuit with an oxygen-free gas is
necessary when working at high temperatures, because high pressure mists
could form an explosive mixture with the oxygen present in the air. Though
there are other thermal oils suitable for slightly higher working temperatures and lower solidification temperatures (e.g. Syltherm 800), they are too
expensive for large solar plants.
The nominal mass flow per row is calculated to achieve a good heat
transfer coefficient between the steel receiver tube and the fluid circulating
inside it, while the pressure drop in the row is kept reasonable. Since the
solar field parasitic electricity consumption depends directly on the pressure drop and the pressure drop in turn depends on the mass flow, the
nominal mass flow per row of PTCs in a solar field is designed as a compromise between a good heat transfer coefficient inside the absorber tubes
and a reasonable pressure drop between the row inlet and outlet. Reynolds
numbers above 105 give a good heat transfer coefficient; additionally, if they
are kept below106 the pressure drop will not be excessive. The number of
PTCs connected in series in each parallel row depends on the nominal mass
flow per row because the higher the flow the more collectors must be connected in series to achieve the nominal temperature difference between the
row outlet and inlet. The reason why several collectors are connected in
series in each row is that a single collector is not able to provide a high
enough temperature difference if the working fluid mass flow is high enough
to guarantee a good heat transfer coefficient. The higher the mass flow the
smaller the temperature difference that can be provided by a single parabolic-trough collector at design point. So the number of collectors to be
connected in series in each row is found by dividing two parameters, the
temperature difference between the solar field inlet and outlet and the
temperature difference that can be provided by a single collector at design
© Woodhead Publishing Limited, 2012
226
Concentrating solar power technology
point. The temperature difference in a single collector at design point is
calculated using the energy balance explained in Section 7.4.2, with the PTC
optical, geometrical and thermal design parameters.
Most of the parabolic-trough collector designs in recent solar thermal
power plants have technical characteristics similar to the EuroTrough-150,
as listed in Table 7.3. Rows that are 600 m long (four 150 m collectors or
six 100 m collectors in series) are becoming common in PTC solar power
plants, because the mass flow of the thermal oil required to achieve a
temperature increase of 100°C in the row meets the heat transfer coefficient and pressure drop recommendations, with a reasonable piping cost.
This configuration requires a design point fluid mass flow per row of about
5 kg/s and a temperature step of 25°C in every PTC. The shape of the plot
of land where the solar field has to be implemented also needs to be taken
into consideration when determining the number of collectors in series per
row.
Once the number of collectors to be connected in series in a row has been
calculated, the next step is to determine the number of rows to be connected
in parallel. This number depends on the thermal power demanded by the
industrial process to be fed. The number of rows is determined by a very
simple procedure: the ratio of the thermal power demanded by the industrial process to be supplied by the solar field to the thermal power delivered
by a single row of collectors at design point. It should be explained here
that the solar fields of solar thermal power plants with the same rated
(nominal) power may be very different sizes, depending on whether they
have a thermal storage system or not. So for instance, a 50 MWe solar power
plant in Spain with a 1 GWh thermal storage system requires a 155-row
solar field with four PTCs in each for a total collecting surface of 510,000 m2,
while a solar plant with the same rated power (50 MWe) and no thermal
storage requires 88 rows with four collectors in each row, with a total solar
field collecting surface of 288,000 m2. The reason for this difference is that
in a solar plant with a thermal storage system, the solar field must not only
supply thermal energy to the power block but also to the thermal storage
system to keep the power block running at full load for seven-and-a-half
hours after sunset. Although both plants have the same rated power, they
have solar fields with very different sizes and their yearly hours of operation
are therefore very different too.
Two parameters are essential for calculation of solar field size and rated
plant power: the solar multiple and the capacity factor. The solar multiple
is the ratio between the solar field thermal output at design point (the
design point is usually set at noon on a summer day in the northern hemisphere) and the thermal power required to feed the power block at nominal
(rated) power. Therefore, the bigger the solar field, the higher the solar
multiple for the same rated power plant. So, in Spain, Morocco and other
© Woodhead Publishing Limited, 2012
Parabolic-trough concentrating solar power (CSP) systems
227
countries at a similar geographic latitude, solar multiples for 50 MWe plants
without a thermal storage system are in the range 1.15–1.30, while for plants
with a 1 GWh thermal storage system, the solar multiple is close to 2.
The capacity factor of the solar plant is the ratio between the number of
equivalent full-load solar-only operating hours a year and the maximum
number of hours of plant operation if it were operated around the clock
(365 × 24 = 8,760 hours). Since a thermal storage system increases the
number of hours of operation, the capacity factor for plants with thermal
storage is higher than plants with none. For instance, in southern Spain, the
capacity factor of a 50 MWe solar plant with a 1 GWh thermal storage
system is about 0.4, while the capacity factor with no thermal storage is
about 0.22.
Since a large commercial parabolic-trough system may have more than
80 km of collectors distributed in many parallel rows, the way in which the
rows are connected and the solar field piping layout are very important to
keep pressure losses, and thus parasitic electricity consumption, reasonably
low. The three basic layouts used in parabolic-trough collector solar fields
are called direct return, reverse return, and center feed, as shown schematically in Fig. 7.17. In every option, the hot outlet piping is shorter than the
cold inlet in order to minimize thermal losses. There are advantages and
(a) Direct return
Collector
Collector
Collector
Collector
Thermal losses
Uneven ΔP
Higher pumping losses
Collector
Collector
Collector
Collector
(b) Reverse return
Collector
Collector
Collector
Collector
Thermal losses
Higher cost
Similar ΔP
(c) Central feed
Shorter piping length
Better access to all the collectors
Uneven ΔP
7.17 Different piping layouts for solar fields with parabolic-trough
collectors.
© Woodhead Publishing Limited, 2012
228
Concentrating solar power technology
disadvantages in each of these three configurations, which are briefly
explained in following paragraphs.
The direct-return piping configuration is the simplest and probably most
extensively used in small solar fields. Its main disadvantage is that there is
a much greater pressure difference between the parallel row inlets, so
manual valves must be installed in them to keep a constant flow through
each row. These valves cause an additional pressure drop in the solar field,
and their contribution to the total pressure loss in the system can be significant. The result is higher parasitic energy consumption than for the reversereturn layout, where the fluid enters the collector array at the opposite end
and the rows with a longer inlet piping have a shorter outlet piping, thus
better balancing the pressure drop associated with each row. However, the
total length of the piping in the reverse-return layout is longer than for the
direct-return configuration, thus increasing thermal losses, although this
strongly depends on the solar field inlet temperature. If this temperature is
low, additional heat loss is negligible. Adding length to the pipes, however,
results in higher piping, insulation, and fluid inventory costs. In a reversereturn piping arrangement, to ensure a completely uniform flow distribution among all the parallel rows without valves, the header pipes must step
down in size on the inlet and step up in size on the outlet to achieve constant
fluid velocity through the headers.
The center-feed configuration is the most widely used layout for large
solar fields. Like the direct-return design, pressure loss in the solar field is
higher if balancing valves are installed at the row inlet. This layout minimizes piping because there is no pipe running the length of the collector
row. Also, direct access to each collector row is possible without any need
for underground piping. This is a significant advantage of the center-feed
layout, because access to the solar field is often required for collector
washing. Since manual valves have to be installed at the inlet and outlet of
every row for maintenance, these valves can be used to balance the pressure
drop in the parallel rows at no extra cost. The center-feed layout has therefore become the preferred option for large PTC solar fields.
It is also very important for solar field piping to be well insulated, because
overall efficiency decreases with inadequate thermal insulation. The length
of the piping is significant and excessive thermal losses would reduce the
amount of useful thermal energy delivered to the power block. Thermal
bridges in piping supports and other components (e.g., oil tanks, pumps,
etc.) must also be avoided for the same reason.
Drain and venting valves in the piping are also important items in the
construction of parabolic-trough collector fields. Venting valves must ensure
that no air bubbles remain inside piping after filling with the thermal oil
working fluid, while the drain valves are essential for maintenance to drain
pipe segments for repair. Since thermal oil pollutes and is flammable at high
© Woodhead Publishing Limited, 2012
Parabolic-trough concentrating solar power (CSP) systems
229
temperatures, no welds or repairs may be done before the pipe segment
involved is fully drained and inertized with oxygen-free gas.
7.5
Operation and maintenance (O&M) of
parabolic-trough systems
The fraction due to operation and maintenance (O&M) of the total cost of
electricity produced by large solar thermal power plants is within the 0.02–
0.035 c/kWh range (approximately 8% of the total cost of the electricity
produced by the solar plant). Since the fuel (i.e., solar radiation) is free
of charge, most of this is manpower for plant operation and system
maintenance.
The most frequent activities related to solar field O&M are periodic
measurement of mirror reflectivity and their washing. Mirror reflectivity
directly affects the amount of useful thermal energy delivered by the solar
collectors, because a 10% higher reflectivity means a 10% increase in useful
thermal energy generated. Dust carried by ambient air is progressively
deposited on the surface of the mirrors gradually reducing their reflectivity
after each mirror washing. Reflectivity of a recently-washed back-silvered
glass mirror is typically 0.93%. Mirror soiling is very site-specific. Experience in Spain is that in summer, reflectivity rapidly decreases at a rate of
about 0.0025% per day during the first two weeks after washing. So in ten
days after washing, reflectivity is only about 0.90, and the mirrors must be
washed again to recover the solar field’s nominal reflectivity and optical
efficiency. Reflectivity decreases more slowly in winter and mirrors need
not be washed as often as in summer. Specially designed mirror washers
are used for this purpose. Demineralized water is carried through the solar
field on a tank truck that pumps it at 200 bar to remove the dust deposited
on the front surfaces of the mirrors. Figure 7.18 (a) shows one of the mirror
washers, called Twister, used at the SEGS plants in California. When mirrors
are not very dirty, simple demineralized water curtains can be used to wash
the mirrors (see Fig. 7.18 (b)). In either case, the consumption of demineralized water required to wash the mirrors is about 0.7 l/m2. Although the
photographs shown in Fig. 7.18 were taken in the morning, mirrors are typically washed at night so the entire solar field remains in operation during
daylight hours for maximum collection and conversion of solar radiation.
Breakage of mirror glass does not occur very frequently in parabolictrough systems (much less than 0.1% per year) and durability of backsilvered glass mirrors is excellent provided that they do not have to withstand
wind speeds over 110 km/h. When the mirror glass breaks, the side effects
can be more important than the breakage itself, because the pieces of falling
glass could hit and break the glass receiver pipe cover, which costs about
800c/unit to replace (plus manpower).
© Woodhead Publishing Limited, 2012
230
Concentrating solar power technology
(a)
(b)
7.18 Typical devices for mirror washing in a parabolic-trough solar
field.
© Woodhead Publishing Limited, 2012
Parabolic-trough concentrating solar power (CSP) systems
231
Another solar field maintenance task is checking collector alignment and
solar tracking systems. Small displacements in the concrete foundations or
malfunction of the solar tracking system can lead to incorrect positioning
of the receiver pipes and significant reduction of the intercept factor.
Periodic maintenance of the ball-joints installed between adjacent collectors to allow thermal expansion of the receiver pipes and independent
collector movement is also required. Their graphite packing must be refilled
every 4–5 years of operation to prevent leaks. Portable chemical detectors
are used to check for the small amounts of oil vapor that always precede a
visible oil leak through the ball-joint packing.
Last but not least, thermal oil parameters and condition must be analyzed
every year. A sample is taken by the plant owner and sent to the oil supplier, who is responsible for the analysis. Although the durability of thermal
oils currently used as heat transfer fluids in parabolic-trough systems is
excellent, the maximum bulk temperature recommended by the supplier
must not be exceeded in order to avoid rapid degradation that would significantly increase the amount of oil that must be replaced yearly. With
proper O&M, less than 3% of the thermal oil has to be replaced every year,
although this percentage could increase up to 20% or higher if the recommended maximum bulk temperature is exceeded often.
7.6
Thermal storage systems
Energy storage is discussed in detail in Chapter 11. Here the specific aspects
as applied to PTC systems are reviewed.
Practical experience with thermal storage systems is more limited than
experience with solar collectors because only the first SEGS plant was
provided with a thermal storage system, which used ESSO 500 thermal oil
in two tanks with a thermal capacity of 140 MWh. This system, which was
put into operation in 1984, was destroyed by a fire in 1999. The operation
of this thermal storage system had been reliable and efficient until that date.
A 5 MWh thermal energy system using 115 m3 of Therminol 55 thermal oil
in a thermocline tank has been in operation at the Plataforma Solar de
Almería (PSA) since 1982, and has proven to be highly reliable with a 0.92
charge/discharge efficiency.
Due to the high environmental risk of large tanks filled with hot thermal
oil, recent large PTC solar thermal power plants have two-tank molten-salt
thermal storage systems with the configuration shown in Fig. 7.10. This
configuration was used in several plants built in Spain during the first
decade of this century. The salt used is a mixture of 60% NaNO3 and 40%
KNO3, with a melting point in the range 225–238°C. With a nominal electrical output of 50 MWe, these plants have a solar field aperture area of
© Woodhead Publishing Limited, 2012
232
Concentrating solar power technology
510,000 m2 and a thermal storage system with a capacity of 1 GWh, which
is able to keep the power block running at full load for seven-and-a-half
hours after sunset. The solar field is big enough to feed the plant power
block and charge the thermal storage system during daylight hours on
clear summer days. The storage system is charged when molten salt pumped
from the cold tank to the hot tank is heated by the oil delivered by the
solar field at 395°C. The system is discharged when the molten salt stored
in the hot tank at 385°C is pumped to the cold tank and its thermal energy
is transferred to the thermal oil, which is then sent to the power block
steam generator. Since the first of these 1 GWh thermal storage systems
with molten salt was put into operation in 2008, practical experience is still
very limited, and a longer time period is needed to assess their durability
and reliability. However, the results are so far very encouraging and
positive.
The predecessor of the 1 GWh molten-salt thermal storage systems now
in use was the molten-salt system tested in the American Solar Two Project
at the end of the last century (James, 2002; Reilly and Kolb, 2001). Since the
size of the storage system tested in Solar Two was only 7% of current
1 GWh systems, the significant difference in size demanded careful analysis
and engineering to solve some technical constraints associated with the size
of the components and melting of 30,000 MT of salt.
The first years of operation with the large 1 GWh molten-salt storage
systems installed in Spain at the end of the first decade of this century have
provided encouraging results. However, the cost of these systems has
increased significantly due to an increase in the cost of salt. The various
other approaches to energy storage that are under investigation are reviewed
in Chapter 11.
7.7
Future trends
Although solar thermal power plants with parabolic-trough collectors are
now profitable in a few countries due to public incentives in the form of
feed-in tariffs or tax credits, it is clear that ways to improve efficiency and
reduce costs must be found, because the current public incentives will be
progressively reduced in the future. The main goal of current incentives is
to make the first commercial projects financially feasible for investors,
thereby stimulating the implementation of first plants. Pushed by the need
to improve the technology and reduce the cost of the electricity generated,
many private and public entities worldwide are carrying out a significant
number of R&D projects to improve components, operation and maintenance procedures, and solar system-to-power block connection.
These R&D activities are also motivated by the growing demand for
parabolic-trough collectors and their components due to public incentives
© Woodhead Publishing Limited, 2012
Parabolic-trough concentrating solar power (CSP) systems
233
implemented in the USA, Spain and other countries at the beginning of the
twenty-first century for electricity produced by solar thermal power plants.
Since these incentives made this type of solar power plant profitable, many
companies undertook development of new PTC designs and components
to meet the growing demand and to lower costs.
Due to the logical limitation of space in this section, it describes only
future trends in working fluids for parabolic-trough collectors and new PTC
designs.
7.7.1
New working fluids
The thermal oils currently used as heat transfer fluids have two clear limitations, their degradation at temperatures above 400°C and the environmental and fire hazards due to possible leakages. The thermal limit imposed
by these oils is a serious barrier to increasing power block efficiency,
because the temperature of the steam delivered to the power block cannot
be higher than 390°C, thereby limiting steam turbine efficiency. However,
since higher working temperatures in the solar field also increase thermal
losses, the overall solar plant efficiency does not increase at the same rate
as the power block. Another advantage of operating the solar field at
higher temperatures is the fact that it decreases the size, and hence the
cost, of the thermal storage system needed to achieve the required storage
capacity.
New fluids are under study for replacing thermal oil: molten salt, pressurized gases and water/steam. All three of these fluids have advantages
and disadvantages when compared to thermal oil, as listed in Table 7.7. The
use of the same molten salt in the solar field and in the thermal storage
system has clear advantages, because the molten salt currently used has
good thermal stability up to 575°C (175°C higher than the oil) and the
overall plant configuration would be simpler, because the oil/molten-salt
heat exchanger now needed in current plants would no longer be needed.
However, the crystallization point of molten salt (>125°C) is significantly
higher than oil (>12°C), and a very efficient, and expensive, heat tracing
system is required in the solar field to avoid solidification of the molten salt
in cold weather.
The use of water/steam for direct steam generation (DSG) would also
avoid the problem associated with thermal oil, but the two-phase flow (i.e.,
liquid water + steam) in the evaporating section of every row of collectors
in the solar field introduces some technical constraints demanding a more
complex solar field control to keep the steam temperature and pressure
stable at the solar field outlet during solar radiation transients. The main
advantage of this option is that plant configuration is so simple, because the
steam demanded by the power block is produced directly in the solar field.
© Woodhead Publishing Limited, 2012
234
Concentrating solar power technology
Table 7.7 Advantages and disadvantages of new working fluids compared to
thermal oil for parabolic-trough collectors
Fluid
Advantages over thermal oil
Disadvantages compared to
thermal oil
Molten salt
• More efficient heat storage
• Higher working
temperature
• No pollution or fire
hazards
Water/steam
• Simple plant design
• Higher working
temperature
• No pollution or fire
hazards
• Higher steam temperature
• Thermal storage
enhancement
• No pollution or fire
hazards
• Higher thermal losses
overnight
• More complex solar field
design
• Higher electricity
consumption
• Lack of suitable storage
system
• More complex solar field
control
• Solar field higher pressure
• Poor heat transfer in the
receiver tubes
• More complex solar field
control
• Solar field higher pressure
Gas
On the other hand, the main barrier at present is the lack of a suitable
thermal storage system for this option, because current storage systems
based on sensible heat are not suitable for DSG solar fields, which delivers
steam that must be condensed to release its thermal energy. Since steam
condensation takes place at a constant temperature, special thermal storage
systems based on latent heat are needed. Such systems must use a storage
medium that melts during charging by steam condensation. The melted
storage medium must crystallize during discharge to produce the steam
required for the power block. Although development of latent-heat storage
systems is already under way, much R&D is still needed before commercial
units become available.
Research in the use of a compressed gas (e.g., CO2, air or N2) inside the
receiver pipes to convert the solar radiation into thermal energy in the form
of the sensible heat of the gas is also under way, because this option would
overcome the barriers associated with the thermal stability and fire hazards
of thermal oil. The possibility of working with gas at temperatures over
500°C is also of great interest because thermal storage would be enhanced
by the greater difference between the hot and cold temperatures, and less
storage medium would be required to store the same amount of energy. The
main constraint on the use of compressed gas is the pressure drop in the
© Woodhead Publishing Limited, 2012
Parabolic-trough concentrating solar power (CSP) systems
235
solar field piping, which would demand more pumping power, and therefore, also more internal consumption of electricity.
Although the technical feasibility of these three new fluids has been
already proven at CIEMAT’s small test facilities in Spain (Zarza et al., 2004;
Rodríguez et al., 2009) and ENEA’s in Italy (Maccari, 2006), the three
options have to be evaluated and tested in pilot plants large enough to
ensure that results can be extrapolated to large commercial plants. A 5 MWe
pilot plant promoted by ENEA with molten salt in the solar field is expected
to enter into operation in 2011, while a 3 MWe plant with direct steam
generation is expected to enter into operation in Spain in 2012 (Zarza
et al., 2008).
7.7.2
New PTC designs
A significant number of new PTC designs were developed in the USA and
Spain in the period 2005–2010, clearly indicating the great commercial
interest in this technology. The new designs were specially aimed at reducing the manufacturing and assembly costs because these items are a significant fraction of the total solar field cost. The use of components made of
stamped sheet steel and the design of special torque tubes with enhanced
resistance to bending are good examples of innovations introduced in
recent PTC designs. A number of these new designs have already been
deployed commercially (e.g., SenerTrough, Skal-ET), while at the end of
2010, others were still awaiting their first implementation in a large commercial plant (e.g., AlbiasaTrough, URSSATrough, etc.).
Future trends in new PTC designs concern two main topics:
•
•
innovative means of providing a stiff structure, other than the torque
box, torque tube or space frame concepts, with a low manufacturing
and assembly cost, while maintaining good mechanical rigidity in wind
loads
Larger collector aperture area and parabola width.
The innovative collector design proposed by the Alcoa Company is a good
example of the first topic, as explained in http://www.alcoa.com/global/en/
innovation/info_page/home.asp. The Alcoa design uses a monolithic structure that enables simple ‘drop-in-place’ collector assembly, and glass reflectors are replaced by highly reflective aluminum mirrors. A first prototype
was installed at NREL’s test facilities in Golden, Colorado (USA) early in
2010 for outdoor testing.
Although most of the new PTC designs retain the general EuroTrough
dimensions (Luepfert et al., 2003), the technical and commercial feasibility
of a parabola width over 5.76 m is currently under evaluation in several
© Woodhead Publishing Limited, 2012
236
Concentrating solar power technology
countries and the first prototypes will be tested in the 2010s. These larger
designs require larger receiver tubes, and sometimes larger parabolic mirror
segments. The main benefit expected from these larger designs is a cost
reduction, especially due to a reduction in the number of drive units and
ball-joints. On the other hand, their main technical constraint is the higher
wind load they have to withstand and the heavier structure required to
support all the components. An example of this trend in parabolic-trough
collectors is the HelioTrough design developed by Solar Millennium, Flagsol
GmbH and Schlaich Bergermann und Partner (SBP) (Germany) with a
parabola width of 6.77 m and a steel receiver pipe diameter of 89 mm. A
first HelioTrough collector loop was under testing at the SEGS V plant in
Kramer Junction (California, USA) in 2010.
Experimental results from outdoor testing with the first prototypes
using new and larger sizes will demonstrate whether these approaches
are a good choice or not. The significant R&D effort undertaken by industry in collaboration with public centers to develop improved collector
designs and components leads us to believe that cost reduction is likely to
be rapid.
7.8
Conclusions
Public incentives in the form of feed-in tariffs or tax credits implemented
in the USA, Spain and other countries during the first decade of the twentyfirst century have promoted a multitude of solar thermal power plant projects, most of them with parabolic-trough collectors, because the successful
track record of the SEGS plants established them as the ‘least risk’ and
most financeable technology. Although the world financial crisis of 2008 was
a serious barrier to development, many of the projects that were initially
proposed at that time have been successful and 19 parabolic-trough CSP
plants with 876 MWe total power output were in operation at the end of
2010.
However, it was clear from the beginning of this new construction phase
that ways to lower costs and increase efficiency had to be found to make
the continuation of commercial deployment compatible with the expected
and logical reduction in public incentives. This was the driving force behind
a significant effort in R&D by industry, engineering firms and public centers
to develop new collectors, components (receivers, solar reflectors) and
fluids (molten salt, direct steam generation and compressed gas) during the
first decade of this century.
Although most of the financial and R&D effort was focused on large-area
parabolic troughs for CSP plants, industrial process heat applications are
also a potential market where smaller parabolic-trough collectors can find
a profitable niche.
© Woodhead Publishing Limited, 2012
Parabolic-trough concentrating solar power (CSP) systems
7.9
237
Sources of further information and advice
Since the commercial deployment of large-scale parabolic-trough systems
is very recent, the generally available literature is not abundant. However,
some documents for further information about various aspects of this technology are given below.
For detailed information about PTC designs developed in the USA in
the 1970s and 1980s see:
Dudley V.E., Workhoven R.M. (1981) ‘Performance testing of the Acurex
Solar Collector Model 3001-03’. Tech. Rep. No. SAND80-0872.
Albuquerque: SANDIA.
Kesselring P., Selvage C.S. (1986) The IEA/SSPS Solar Thermal Power
Plants. Vol. 2: Distributed Collector System (DCS), 1st edn, Berlin:
Springer-Verlag.
Dudley V.E., Workhoven R.M. (1982) ‘Performance testing of the Solar
Kinetics T-700A Solar Collecto’. Tech. Rep. No. SAND81–0984.
Albuquerque: SANDIA.
Cameron C.P., Dudley V.E. (1986) ‘Solar kinetics incorporated modular
industrial solar retrofit qualification test result’. Tech. Rep. No. SAND85–
2320. Albuquerque: SANDIA.
Cameron C.P., Dudley V.E., Lewandowski A.A. (1986) ‘Foster Wheeler
solar development corporation modular industrial solar retrofit qualification test results’. Tech. Rep. No. SAND85–2319. Albuquerque: SANDIA.
Concerning basic principles of optics and geometry, two books with useful
information are:
Rabl A. (1985) Active Solar Collectors and their Applications. Oxford:
Oxford University Press.
Duffie J.A., Beckman W.A. (1991) Solar Engineering of Thermal Processes.
New York: John Wiley & Sons.
A document published by Sandia National Laboratories in 1999 on the
experience gained at the SEGS plants is very valuable for operation and
maintenance of parabolic-trough systems:
Cohen G.E., Kearney D.W., Kolb G.J. (1999) ‘Final report on the operation
and maintenance improvement program for concentrating solar power
plants’. Tech. Rep. No. SAND99-1290. Albuquerque, SANDIA.
In 2010 the International Energy Agency published the book entitled, Technology Roadmap for Concentrating Solar Power, available at hhtp://www.iea.
org. This document gives a complete overview of the commercial situation of
solar thermal power plants, the expected cost reduction, and boundary conditions required for large commercial deployment in Sunbelt countries.
© Woodhead Publishing Limited, 2012
238
Concentrating solar power technology
For further technical information related to R&D activities as well as
commercial and financial information, the Proceedings of the SolarPACES
conferences held in Berlin (Germany), Perpignan (France) and Granada
(Spain) in 2009, 2010 and 2012, respectively, contain very interesting papers.
The plenary sessions of these conferences are especially interesting for
subjects related to strategy and market.
For specific information on R&D activities, the web pages of public
research centers usually have yearly reports available in PDF format. For
example, the annual reports of the Plataforma Solar de Almería (Spain) are
available free of charge at: http://www.psa.es.
7.10
References and further reading
Fernandez-García A., Zarza E., Valenzuela L. and Pérez M. (2010) ‘Parabolic-trough
solar collectors and their applications’. Journal of Renewable and Sustainable
Energy Reviews, 4(7), 1695–1721.
Gunther H. (1922) Technische Traeume (Technical Dreams), Rascher und Cie.
Verlag, Switzerland.
Harats Y. and Kearney D. (1989) ‘Advances in Parabolic-Trough Technology at the
SEGS Plants’. Proceedings of the 1989 ASME International Solar Energy Conference, San Diego, April, 471–476.
James E.P. (2002) ‘Final Test and Evaluation Results from the Solar TWO Project’.
Report SAND2001-0120, Sandia National Laboratories, Albuquerque.
Lotker M. (1991) ‘Barriers to Commercialization of Large Scale Solar Electricity.
The LUZ Experience’. Tech. Rep. No. SAND91-7014, Sandia National Laboratories, Albuquerque.
Luepfert E., Zarza E., Schiel W., Osuna, R., Esteban A., Geyer M., Nava P., Langenkamp J. and Mandelberg E. (2003) ‘Eurotrough collector qualification complete
– Performance test results from PSA’. Proceedings of the ISES 2003 Solar World
Congress, June 16–19, Göteborg (Sweden).
Maccari A. (2006) ‘Innovative heat transfer concepts in concentrating solar fields’.
Available at: http://ec.europa.eu/energy/res/events/doc/maccari_enea.pdf.
Price H., Lüpfert E., Kearney D., Zarza, E., Cohen Gee R. and Mahoney R. (2002)
‘Advances in parabolic-trough solar power technology’. Journal of Solar Energy
Engineering, 124, 109–125.
Pytlinski J.T. (1978) ‘Solar energy installations for pumping irrigation water’. Solar
Energy, 21, 255–258.
Reilly H.E. and Kolb G.J. (2001) ‘An Evaluation of Molten-salt Power Towers
Including Results of the Solar Two Project’. Report SAND2001-3674, Sandia
National Laboratories, Albuquerque.
Rodríguez M.M., Marquez J.M., Biencinto M., Adler J.P. and Díez L.E. (2009) ‘First
experimental results of a solar PTC facility using pressurized gas as the heat
transfer fluid’. Proceedings of the SolarPACES-2009 Congress, Berlin, September
15–18.
Shaner W.W. and Duff W.S. (1979) ‘Solar thermal electric power systems: comparison of line-focus collectors’. Solar Energy, 22, 13–49.
© Woodhead Publishing Limited, 2012
Parabolic-trough concentrating solar power (CSP) systems
239
Zarza E., Valenzuela L., León J., Hennecke K., Eck M., and Weyers H.-D. (2004)
‘Direct steam generation in parabolic troughs: Final results and conclusions of
the DISS project’. Energy – The International Journal, 29(5–6), 635–644.
Zarza E., López C., Cámara A., Burgaleta J.I., Martín J.C. and Fresneda A. (2008)
‘Almería GDV – The First Solar Power Plant with Direct Steam Generation’.
Proceedings of the 14th SolarPACES International Symposium, Las Vegas, March
4–7.
© Woodhead Publishing Limited, 2012
8
Central tower concentrating solar power
(CSP) systems
L. L. VA N T- H U L L, formerly University of Houston, USA
Abstract: In this chapter we first address the conception, design and
construction of central receiver tower systems, including a summary of
commercial plants operating or in construction. We then discuss a variety
of issues affecting the design and performance of central receiver
systems. These include initial considerations, elements of cost and
performance, characteristics of the heliostat, characteristics of the
receiver, and any external constraints on the design such as flux density
limitations or land constraints. Finally, several variants on the simple
configuration are critically discussed.
Key words: solar central receiver, solar power tower, concentration
optics, optimization, heliostat field, beam errors, spillage, constrained
optima, off-axis aberrations, secondary reflectors, beam down, receiver
flux density, field layout, net energy.
8.1
Introduction
A central receiver system consists of an array of tracking mirrors, or heliostats, which are spaced in a field to avoid mechanical or optical interference
with one another as they pivot to reflect incident direct-beam sunlight onto
an elevated receiver or secondary reflector (Hildebrand and Vant-Hull,
1977). The receiver is designed to effectively intercept the concentrated
incoming sunlight (solar energy) and (usually) absorb it as heat at an elevated temperature. This energy is collected by a working fluid and stored
as thermal energy, used to drive an electrical generator, or used as process
heat. Many of the additional issues which must be addressed in designing,
building, and operating a complete solar thermal power station are discussed in more detail in a recent Sandia report (Kolb, 2011). Alternatively,
photovoltaic panels could replace the thermal receiver, or the light (photons)
can be used directly to drive a chemical reaction or even a laser. The optical
design and optimization of central receiver (CR) systems (also known as
solar power towers) are somewhat complicated by the multitude of variables one must consider and the continuous variation in configuration and
performance of each of the heliostats as they track the sun and interact with
one another. However, the efficient collection, high concentration and high
temperature of the heat collected are of interest for many applications.
240
© Woodhead Publishing Limited, 2012
Central tower concentrating solar power (CSP) systems
241
Central receivers have the advantage that all the solar energy conversion
takes place at a single fixed region, i.e., the receiver. This allows the receiver
to be fixed, largely avoiding the need for energy transport networks, and
allows more cost-effective investment designed to improve the efficiency
and sophistication of the energy conversion process. They have been built
most often as single large systems to power a steam cycle; however, smaller
systems or modular systems employing multiple towers have attractions in
some applications. A universal disadvantage is that the fixed position of the
receiver means that heliostats do not generally point directly at the sun, so
that the amount of collected solar radiation per unit area of mirror is
reduced compared to a dish concentrator (the cosine effect). Of course,
much of the reflector in parabolic dishes or troughs is also tilted with respect
to the sun so the reflector area is, again, larger than its aperture area.
However, in these collectors, little or none of the incoming light is bypassed
to the ground, as is characteristic in Fresnel systems such as the solar tower
or linear Fresnel systems. In such systems, facets of the parabolic surface
are projected to the ground where the tracking mirrors redirect the sun to
the receiver. Consequently, it is necessary to forego collection of some of
the incoming energy by spacing the mirrors in order to avoid serious shading
of adjacent mirrors as the sun moves, or blocking of some of the reflected
light on its way to the receiver. These issues result in a trade-off among the
competing events in order to collect light onto the receiver most cost effectively. This chapter addresses the conception, design and construction of
central receiver tower systems and a variety of issues affecting the design
and performance. Further consideration of heliostats and their size/cost
optimization can be found in Chapter 17.
8.1.1
Basic configurations
The CR concept can be realized in several configurations, defined essentially by the receiver, as indicated in Fig. 8.1. If the receiver consists of an
external cylinder, the absorbing surface can be seen from all directions,
resulting in a surrounding field of heliostats defined primarily by their relative efficiency in directing sunlight onto the receiver (Fig. 8.1(a)). For a
given power level, this results in a shorter and lower cost tower and vertical
piping. The associated disadvantage is that the heated surface is exposed to
the elements and all thermal re-radiation and convection is lost.
The principal alternative is a cavity receiver in which the heated surface
is contained in an insulated enclosure containing a large aperture to admit
the sunlight (Fig. 8.1(b) and (c)). This has the effect that light can only be
effectively collected from heliostats within a cone defined by the normal to
the aperture, as heliostats far off-normal will see an aperture foreshortened
by the cosine of the cone angle (50% at 60°). Usually the aperture is directed
© Woodhead Publishing Limited, 2012
242
Concentrating solar power technology
(a) McDONNELL DOUGLAS
TOWER EL. 86 m
COLLECTOR FIELD AREA – 29.5 ha
HELIOSTATS – 1760 (McDONNELL DOUGLAS)
1643 (BOEING)
368.3 m
(b) HONEYWELL
TOWER EL. 159 m
COLLECTOR FIELD AREA – 20.6 ha
HELIOSTATS – 1598
401.25 m
TOWER EL. 122 m
(c) MARTIN MARIETTA
COLLECTOR FIELD AREA – 33.7 ha
HELIOSTATS – 1555
652 m
(d)
Virtual object
f1 = 1 unit
Hyperbolic
secondary
Beam
divergence
(exagerated)
CPC or
receiver
aperture
Plane of heliostat axis
f2 = 3 units
Real
image
linear magnification
f
= 2 = 3 times
f1
8.1 (a), (b) and (c) To-scale sketches of CR configurations proposed for
the Solar One facility: external receiver with surround field, downlooking cavity receiver with surround field, cavity receiver with north
field (modified from Sandia sketches); (d) also a representative beamdown surround field with secondary hyperbola (sketch by author).
© Woodhead Publishing Limited, 2012
Central tower concentrating solar power (CSP) systems
243
somewhat downward and toward the pole to face an array of heliostats
positioned to most effectively illuminate the aperture as in Fig. 8.1(c). This
means the heliostats will tend to be primarily on the polar side of the aperture (north field in the northern hemisphere, south field in the southern
hemisphere), where the angle of incidence on the heliostats at a noon time
design point is best. The resulting field will be more optically efficient than
a surround field around noon, but less efficient in the morning and afternoon
when the cosine effect on the heliostats plays a large role. With a noon time
design point, but for the same heliostat area, the annual energy collected will
be lower compared to the surround field. This is because the west field performs very well in the morning and the east field does well in the afternoon
when the polar field performance suffers large cosine losses.
Cavity receivers are most likely to play a role where the required output
temperature is very high (of the order of >1,000 K), or the allowable flux
density on the receiver surface is very low (as for a gas cooled tubular
receiver). Although it is commonly thought the cavity receiver will have
significantly less losses than an external receiver, as we discuss in Section
8.6.2 this is not always the case. Because the cavity can only view a limited
section of the surrounding terrain, its tower will be taller to collect the same
amount of energy. Alternatively, several separate cavities viewing different
fields (e.g., NE and NW, or E, N, W) can be mounted on a single tower, or
the aperture can be horizontal as in Fig. 8.1(b). Partial (or shallow) cavities
are also used.
A third alternative is the beam-down concept. Here all the heliostats direct
their beams at a point (the virtual focus) but a (generally) hyperbolic secondary mirror intercepts the light and redirects it toward the ground (Fig. 8.1(d)).
Here, at the lower focus of the secondary, the redirected sunlight is captured
by an upward-looking receiver often supplemented by a compound parabolic concentrator (CPC) to recover some of the concentration. This is
important because the secondary magnifies the image which would have
been formed at the initial focal point by the ratio of the distance from the
vertex of the secondary to each of the two focal points (the linear magnification, LM). If one wishes to make the secondary small, it must be placed near
the virtual focal point to intercept all the light from the heliostat field.
However, this results in a large LM of the final image and reduces the concentration by the square of the LM. To achieve a small LM requires a large
secondary which is more costly and will create additional shading of the field.
8.2
History of central receivers
8.2.1
Early evolution
Delivery of solar energy as thermodynamically useful heat for power cycles
requires temperatures significantly above the 100°C or so available from
© Woodhead Publishing Limited, 2012
244
Concentrating solar power technology
flat plate collectors, so tracking concentrating collectors are required. To
deliver significant commercial quantities of solar energy, either as electricity
or process heat, total collector areas might typically be of the order of a
square kilometer, which clearly cannot track the sun as a monolithic structure. One solution is to combine a multitude of point focus (dish) collectors,
or line focus (trough) collectors to achieve the required power level. This
involves a multitude of distributed receivers and an energy collection and
transfer network to assemble the collected energy for use.
An alternate solution, reported by the Russian, Victor Baum (Baum
et al., 1957), is to conceptually break the reflector into facets and project
them outward from the receiver to the ground. These ground-based segments (heliostats) can be individually tracked to maintain the reflected
direct-beam sunlight on a receiver. As heliostats on the polar side of the
receiver have an average incidence angle closer to zero than those on the
equatorial side, they are more effective. Consequently, Professor Baum
proposed to mount multiple heliostats on trains of rail cars, which would
move on a set of circular rails to predominantly stay on the polar side of
the receiver, in order to maintain a constant geometry vs. the sun’s azimuthal motion during each day. The elevated cavity or flat-plate receiver
at the center of the circles would rotate in synchronism to face the array,
from west to north (in Russia) to east in order to capture the reflected rays.
He received permission to build the device at his university, but no funding
was provided (private communication).
In Italy (San Ilario), Professor Giovanni Francia (1968) reported on a
‘receiver oriented’ drive mechanism, which automatically redirected a
solar beam to the receiver when activated by a simple clockwork drive.
Periodic (≈weekly) declination adjustments were also required. The
∼130 kWt system of 135 m2 of ≈1 m diameter focused mirrors, built in 1968,
was actuated by a pendulum clock driven by weights, and worked quite well
for over a decade, producing very high temperature steam with little
supervision.
In the 1950s Professor Felix Trombe at Odellio, France, built several solar
furnaces using a single large tracked mirror to fill a fixed horizontally oriented parabola with paraxial sunlight. The largest of these was a 1 MWt
solar furnace (Trombe, 1957). He installed the ≈2,000 m2 faceted parabola
as the north side of an eight-story laboratory, with a focal building about
30 m in front of it. He then broke the required 2,835 m2 primary mirror
required to ‘fill’ this parabola into 63 flat heliostats, each of 45 m2, which he
distributed carefully on the steep side of an adjacent mountain to provide
parallel horizontal beams of sunlight to fill the parabola for several hours
each day, as shown in Fig. 8.2. The highly flat and precisely tracked heliostats
and the ∼9,000 precisely adjusted curved facets of the parabola produced
excellent results, with a peak flux density of ∼13,500 suns. These were the
© Woodhead Publishing Limited, 2012
Central tower concentrating solar power (CSP) systems
245
8.2 The French 1 MW solar furnace at Odellio employing a field of 63
heliostats each 45 m2 in area. The working area is in the focal building
just in front of the large parabolic mirror.
first ‘commercial’ heliostats and they remain representative of the current
generation.
In 1972, Hildebrand et al. published a ‘reinvention’ of the central receiver
as a hemispherical or cylindrical receiver atop a tall tower surrounded by
a field of carefully positioned heliostats. Thinking big, they evaluated a
conceptual 450 m tower producing 565 MWth at 1,000–2,000 K (to power a
steam turbine or a magneto hydrodynamic generator). No technical impediments were envisaged with the tower or heliostat field. (The proposed tower
was about twice as high as current analysis shows, is actually required to
deliver 565 MWth.)
8.2.2
International test facilities and pilot plants
Between 1980 and 1990, a number of international test facilities and pilot
plants were built and operated as shown in Table 8.1. Because of their small
scale, all of these facilities (other than Solar One and Two) employed ‘north
field’ designs (for northern hemisphere) using a flat panel or small cavity
receiver. Only in the range above about 10 MWe do a surround field and a
cylindrical receiver begin to become economically advantageous due to the
reduced tower and piping costs and to the complications implicit with a
single large cavity receiver.
These facilities provided experience with various types of heliostats and
various working fluids, including oil, water steam, molten salt, and sodium.
Most of them also employed small storage units and were equipped to
generate electricity, although not as a commercial entity. Many lessons were
© Woodhead Publishing Limited, 2012
© Woodhead Publishing Limited, 2012
Water/
steam
Heat
transfer
fluid
Storage
capacity
(MWhth)
or heat to
provide
…
Period of
service
1981–84
1981–85
1.0 MWhe
Sodium
Adapted from Lovegrove and Luzzi (2002).
1980–81
3 MWhe
Cone/
Cavity
Water/
steam
Cavity
Receiver
0.036 MWhth
∼2 ha
43 m
∼2 ha
69 m
3.5 ha
55 m
Cavity
3,655 m2
12,912 m2
6,260 m2
7,920
750
1982–88
28 MWhe
External
cylinder
Water/
steam
29.1 ha
80 m
71,095 m2
950
1996–99
107 MWhth
External
cylinder
Molten
salt
35 ha
80 m
81,344 m2
950
56 MWt
43.4 MWt
7.7 MWt
5.95 MWt
10 MWe
10 MWe
0.5 MWe
1 MWe
1 MWe
850
Net turbine
rating
Thermal
power
(MWt)
Irradiance
(W/m2)
Reflector
area
Field area
Target
height
Solar Two
(USA)
Solar One
(USA)
IEA-CRS
(Spain)
Sunshine
(Japan)
Eurelios
(Italy)
Table 8.1 Summary of central receiver demonstration electric power plants
1983–84
3.5 MWhe
Water/
steam
Cavity
7.7 ha
60 m
11,880 m2
700
7.7 MWt
1.2 MWe
CESA 1
(Spain)
Cavity
61 m
7,845 m2
1983–86
12.5 MWhe
1984–85
2.5 MWhe
1985–
1.5 MWhe
Water/
steam
External
80 m
40,584 m2
800
∼1,000
2001–
n.a.
CPC + cavity
‘porcupine’
Beam down
air
80 m
3,500 m2
0.5 MWe
∼
5 MWt
Weizmann
(Israel)
SES 5
(CISUSSR)
5.5 MWt
0.75 MWe
MSEE
(USA)
Molten salt Molten
salt
Cavity
∼2 ha
106 m
10,740 m2
1,040
8.9 MWt
2.5 MWe
Themis
(France)
Central tower concentrating solar power (CSP) systems
247
learned from these facilities concerning control systems, heliostats, pumps,
valves, receivers and working fluids. One of the most important was that
great care must be exercised in selecting ‘off the shelf’ commercially produced items. Often they are designed for a few hundred or thousand thermal
stress cycles during their projected lifetime, but solar plants cycle once or
several times a day, and the ‘life’ can be used up in a year or two, resulting
in failure.
8.2.3
Solar one and solar two
The central receiver system built in the Mojave desert in 1981–2 deserves
special consideration. This system represented a major milestone as the first
system in the world configured as a mature pre-commercial pilot plant. It
started life as ‘Solar One’ and was later re-configured and re-launched as
‘Solar Two’, shown in Fig. 8.3. Lessons from this system are central to, and
are still being utilized by, the present commercial activity with tower systems.
In 1973, Vant-Hull and Hildebrand at the University of Houston received
funding from the National Science Foundation (under the Research Applied
to National Needs program) to study the feasibility of ‘Solar Thermal Power
8.3 Solar Two, the molten-salt system with 28 MWhe of two tank
(warm and hot) salt storage: using the tower and heliostat field of
1,818 heliostats each of 39 m2 of Solar One and 10,000 m2 of
additional larger heliostats (Boeing).
© Woodhead Publishing Limited, 2012
248
Concentrating solar power technology
Systems Based on Optical Transmission’. This study led directly to the construction of Solar One, a 10 MWe ‘pilot plant’ tied to the Southern California Edison grid. This facility was a scaled down prototype of an optimized
100 MWe central receiver system using steam as a working fluid and designed
for commercial operation. Three teams with significantly different design
concepts (shown in Fig. 8.1) competed in a US Department of Energy
funded design study for the pilot plant: McDonnell Douglas (using pedestal
heliostats surrounding a central cylindrical receiver), Honeywell (using
heliostats consisting of four elevation-tracked mirrors mounted on a rotating frame to illuminate a down-facing cavity receiver), and Martin Marietta
(using U-frame type pedestal heliostats placed on the polar side of a polefacing cavity receiver). In the event (DoE, 1977), Solar One used a 45 MWt
cylindrical receiver (McDonnell Douglas, provided by Rocketdyne) at the
focal point (76 m) of a surround field of 1,818 pedestal mounted glass/metal
heliostats (McDonnell Douglas design, provided by Martin Marietta).
These nearly square heliostats had an area of 39 m2 consisting of 12 facets,
each nominally focused and canted to superimpose the solar images from
all facets of the more distant heliostats at the receiver.
The optical design process for the heliostat field, tower and receiver
sought to minimize the capital cost plus present value of operations and
maintenance divided by the annual thermal energy delivered to the ground.
It closely followed the methods and considerations discussed in Section 8.4.
An optimal field layout was established to achieve the minimal value for
the levelized cost of heat delivered to the ground for a hypothetical commercial 100 MWe system. In an iterative process, the focal height, receiver
dimensions, and heliostat field were varied (subject to flux density limitations on the receiver which were ameliorated by an automated aiming
strategy). For Solar One, these results were not re-optimized, but were
scaled to 10 MWe so the pilot plant would better emulate the issues that
would arise in a commercial facility.
Solar One utilized a ‘once through to superheated steam’ receiver, so the
conservative allowable flux on the receiver of this pilot plant was only about
300 kW/m2, consequently the receiver was rather large (13.7 m high and
7 m diameter to handle 45 MWt maximum absorbed energy). In the event,
budget constraints led to a decision to remove some of the heliostats which
were least efficient at the equinox noon design point (on the south side of
the field, poor cosine, etc.); consequently, the south part of the receiver
could not generate 510°C steam and the south quadrant was relegated to a
water preheat function.
In Solar One, storage was provided in the form of a single rocks- sandand oil- thermo-cline tank which was sized to allow 4 hours of turbine
operation at 7 MWe, i.e. 145 MWht. At ground level, the 440°C superheated
steam was split between the turbine inlet and a steam/oil heat exchanger.
© Woodhead Publishing Limited, 2012
Central tower concentrating solar power (CSP) systems
249
Caloria HT-43 from the heat exchanger was delivered to the top of the
storage unit at 304°C to charge the unit, and upon demand, extracted from
the top at 302°C to a steam generator delivering steam to the turbine at
276°C. While the storage unit operated satisfactorily, it was primarily used
for testing purposes and to deliver auxiliary steam for start-up and during
cloud events. As operation under storage steam was only about 70% as
efficient as normal due to the lower steam temperature (277°C vs. 500°C),
and operating the storage system was cumbersome, it was not normally used
to drive the turbine in this test facility.
Overall, Solar One met most of its test objectives in the initial two-year
test period (1982–84), and operated for an additional three years (1984–87)
in a power production mode until support funding for the pilot plant ran
out. During these five years, many useful lessons were learned for future
plants (Kelly, 2000). Generation of steam in the receiver at over 500°C and
high pressure required the use of very heavy wall tubing, which was subject
to large heat fluxes. The thermal gradients, and resulting stresses significantly limited the allowable flux density. In addition, rapid changes in
receiver power resulting from cloud passage, etc., made control difficult, and
subjected the turbine to rapid changes. This was alleviated somewhat by the
stratified-bed storage system, but that was also difficult to use, and direct
storage of steam at high temperature and pressure is costly at other than
small scale. Such considerations led to a search for a single phase medium
for use in the receiver which could also be used directly for storage.
Due to its high specific heat and boiling point and its good heat transfer
properties, non-toxicity, and modest cost, it was determined that (Na60%,
K40%)NO3 salt would provide a superior receiver working fluid, heat transfer fluid, and storage fluid, even compared to pure liquid sodium, which has
better heat transfer properties, but is more difficult to handle and more
expensive (Utility Study, 1988). These conclusions were supported by the
results from several of the small-scale test pilot plants operated in the 1980s.
Consequently, the heliostat field, tower, and turbine of Solar One were
reconfigured as Solar Two. The receiver was replaced with a molten salt
receiver (with approximately three times the allowable flux density and
around one-third the active area), the receiver feedwater pumps and piping
were replaced to allow salt transport, and the storage unit was replaced with
warm and hot salt tanks and a salt-to-steam generator. The room of 10-yearold computers was also replaced with a pair of briefcase-sized DEC Alphas,
and 10,000 m2 of new 95 m2 heliostats were added, primarily on the south
side to overcome the problems of uneven receiver illumination noted at
Solar One.
In the design of Solar Two, the early decision to direct all hot salt from
the receiver to the hot salt tank and to only operate the turbine from steam
generated using salt from storage (Fig. 8.4) made the collection and the
© Woodhead Publishing Limited, 2012
250
Concentrating solar power technology
Receiver
566 °C
Salt
Tower
Heliostat
Salt
Cold
salt
storage
tank
Hot
salt
storage
tank
288 °C
Salt
Turbine
generator
Steam
generator
Feedwater
Steam
Steam
Heat rejection
8.4 Schematic of Solar Two operation showing two tank (warm and
hot) molten-salt storage of the receiver heat transfer fluid (Bechtel
Group International, with permission).
dispatch of energy completely independent of each other. If there was hot
salt and demand, the turbine could be operated; if there was warm salt and
sunlight, the receiver could be operated, making the collected energy fully
dispatchable.
The low pressure of the salt compared to steam allowed much thinner
walled receiver tubes to be used at Solar Two compared to Solar One. This
reduced thermal stress, and along with the much better heat transfer characteristics of the salt, allowed a much higher receiver allowable flux density,
up to 1 MW/m2 (and in future designs up to 1.5 MW/m2). Combined with
a multilevel vertical aim strategy, this allows use of a much smaller receiver
than used in Solar One.
Because the receiver is drained every night to avoid freezing of the salt
(freezing point 220°C), it must be preheated each morning prior to filling,
to avoid tube blockages – the heliostat field is used for this purpose. Only
about 10% of the heliostats were used to avoid overheating of the empty
tubes, and these heliostats must be selected primarily from the sunrise side
of the field (where the cosine efficiency is very low) to provide a uniform
heating. A special processor, which sequentially identifies the heliostat contributing most to any computed hot spot and removes it from track, was
found to be quite satisfactory for this purpose (Vant-Hull et al., 1996). Due
to the high velocity of salt flow required to achieve the high heat transfer,
the salt flow is multi-pass. Salt enters on the (high flux density) polar side,
flows in serpentine fashion to the east and west, crosses over to balance
the diurnal difference between the power delivered by the east and west
fields, and exits from the equatorial-side panels. Here, the lower flux density
© Woodhead Publishing Limited, 2012
Central tower concentrating solar power (CSP) systems
251
accommodates the lower heat transfer coefficient of the hot salt and its
propensity, at the 565°C outlet temperature, to initiate corrosion in contact
with the hot tube (Bradshaw, 1987; Smith, 1992). In fact, it was found to
be cost effective to move some heliostats from the high-performance north
field to the lower-performance south field. This reduced the total power
on the limiting northern panels, and the resulting lower power level and
flux density there allowed use of a shorter and lower cost receiver while
remaining within the flux density limitation. The 565°C salt in the downcomer was depressurized by flow restrictors and deposited in the hot salt
tank at atmospheric pressure. Upon demand, hot salt was pumped through
a preheater-boiler-superheater heat exchanger train, and the resulting
superheated steam was directed into the same 10 MWe turbine used for
Solar One.
Several easily-preventable failures reduced the operating time for Solar
Two. The prefabricated heat trace tape at the top end of the salt riser was
found by the installers to be too long (due to use of too large a pitch while
winding), so was double wound at the top to use the whole length and avoid
field installation of new connectors. Consequently, the top of the riser pipe
became overheated and began to oxidize to the extent that, after a time,
rust particles broke off and eventually caused a number of receiver tube
failures due to blockage. In addition, a portion of the riser pipe had to be
replaced. Better quality control during construction or an appropriate filter
would have prevented this costly exercise.
After a time, the steam generator failed. It turned out that it was designed
using a common utility steam generator code. Inadequate salt circulation
in a localized area resulted in excessive stress during each thermal cycle,
and failure resulted after 500 cycles or so. It should be noted that 500
thermal cycles would actually represent a very long life for a utility steam
generator under normal operating conditions, highlighting the challenges
of using existing commercial solutions in CSP systems. It had to be removed,
redesigned, and rebuilt.
After two years of test and three years of grid tied operation, the receiver
panels developed considerable warpage. This was initially due to inadequate allowance for thermal expansion in the original design, which assumed
normal operating conditions for the panels. During off-design conditions,
the entire length of the panels could be at the maximum temperature and
so experience constraint and warpage. Once this constraint was remedied,
residual flux gradients caused significant additional warpage by the time
operation was terminated.
Several ‘peripheral’ problems such as this depleted the operating budget
to the extent that, shortly after the pre-designated test plan had been completed, Solar Two was shut down (in 1999) prior to significant commercial
operation. Thus, much of the operating experience so useful in establishing
© Woodhead Publishing Limited, 2012
252
Concentrating solar power technology
the ‘bankability’ of the molten salt central receiver concept did not
eventuate.
8.2.4
Period of transition
The period from 1990 to 2005 was a difficult time for solar energy in general,
and due to the large capital investment required and the restricted years of
operation of demonstration plants, for central receiver plants in particular.
During this period, substantial progress was made in heliostat design leading
to lower costs and higher performance. In addition, receiver coatings were
improved, and work was initiated to develop specialized molten salts with
lower freezing points and higher working temperatures. Studies using fluids
other than steam or molten salt in the receiver were undertaken. In particular, supercritical steam or supercritical carbon dioxide offer significant
advantages in turbine generator efficiency. However, the very high temperatures and pressures involved result in very high costs for the receiver, and
even for the risers and downcomers. The containment costs of supercritical
fluids at very high temperatures and pressures make storage of the working
fluid itself expensive (Kelly, 2010).
An alternative receiver concept is a volumetric receiver in which ambient
air is drawn through a ‘thick’ porous mesh or grid, which is simultaneously
irradiated at high flux density. Volumetric receivers can be operated at
either atmospheric or elevated pressure. If the latter, then a transparent
window is required to maintain the pressure while allowing radiation onto
the absorber surface. A window also allows the use of an alternative working
fluid. Extensive experimental and development work was carried out on
the volumetric approach by researchers in Europe and Israel. The incident
flux is absorbed within the volume of the mesh, heating it. Meanwhile, the
airflow effectively cools the outer surface of the mesh, thereby reducing its
temperature and re-radiation. The air is further heated within the hot mesh
and by the ambient solar and infrared radiation. In this way, sufficiently
high temperatures (>900°C) can be achieved to run an efficient Brayton
cycle engine, with storage possible by passing the air through a ceramic bed.
In the years since 1970, there was increasing awareness of the issues of
global warming, peak oil and the limitations of fossil fuel reserves, and the
general issue of the uncosted effects of pollution from fossil fuels. Consequently, starting about 2000, several organizations renewed efforts to implement central receiver power plants. Many studies were undertaken for
specific applications and sites, but for a long time economic conditions were
not appropriate to allow successful financing of the first plant, even with
substantial governmental incentives or tax breaks. For example, Bechtel
(Nexant) carried out a preliminary design for a plant in Spain (Solar Tres,
2000) in response to the Royal Decree promising significant feed-in tariffs,
© Woodhead Publishing Limited, 2012
Central tower concentrating solar power (CSP) systems
253
and in South Africa (Eskom, 2002) in response to a request from the state
utility, Eskom, but in neither case was it possible to close the financial deal.
Certainly, other organizations had similar experiences.
Several companies and organizations developed experimental/test facilities at this time to better position themselves and their new concepts for
the commercial boom that was expected as economic, political, and energy
requirements developed.
8.3
Activities since 2005
8.3.1
Research, development and demonstration
As with CSP in general, 2005 marked a change in activity for central receivers and the beginning of the contemporary period of industry expansion.
•
•
•
•
•
•
•
•
The test facilities at Sandia and at Plataforma Solar were upgraded,
providing facilities for the testing and verification of small-scale
projects.
The Themis site has been designated to test 4 MWt volumetric gas
receivers in an open cavity.
Abengoa Solar New Technologies have been operating a 5 MW research
tower in between its PS10 and PS20 towers. A fourth small tower is
under construction at the same facility for development of air
receivers.
CSIRO in Australia built two small high flux towers (600 kWth and
1.2 MWth) for development of high temperature concepts in steam,
storage, solar fuels, and both air and supercritical CO2 Brayton cycles.
They operate regularly at >1,000 K and up to 1,600 K. The larger of the
two towers uses 450 small (4.5 m2) low-cost, high precision heliostats
(Stein, 2011).
At Jülich, Germany, a north field solar tower facility was designed to
test open volumetric air receivers at the 1.5 MWe level. It has been
operating up to 650°C since March 2009.
Mitaka Kohki of Japan has built a small experimental scale beam-down
system with an elliptical secondary at 10 m elevation illuminating a
cavity equipped with a CPC to recover some of the concentration.
Seventy highly precise heliostats carrying four each 1/4 m2 facets power
the system.
The Chinese Academy of Sciences has built a 7.5 MWth experimental/
demonstration plant 75 km north of Beijing. This plant uses 100 m2
heliostats to illuminate a cavity receiver producing steam and includes
a storage unit (Fig. 8.5).
Korean Institute of Energy Research have built a 200 kWe tower in
Daegu, Korea.
© Woodhead Publishing Limited, 2012
254
Concentrating solar power technology
8.5 Photo of the 7.5 MWth experimental/pilot plant of the Chinese
Academy of Science near Beijing, China. The illuminated initial cavity
receiver/tower stands behind the new tower which has several test
apertures (Zhifeng Wang, with permission).
Several companies were interested in larger scale commercial facilities.
Test platforms were built by companies including Abengoa Solar, LuzII
(BrightSource) and a number of smaller companies to qualify their proprietary heliostats, control systems, and receivers.
In Israel, a number of the engineering and management personnel who
designed and built the LUZ parabolic trough plants in the 1980s (345 MWe
still operational in 2012) reconstituted themselves as LUZII and initiated
the design of a central receiver system. Their first effort was to produce a
20 MWth heliostat field, control system, and receiver test facility in the
Negev desert in Israel. The BrightSource (LuzII) test facility in Israel really
approximates a sector of their first plant and consists of 1,600 heliostats of
14 m2 in two facets, illuminating a flat panel receiver atop a 60 m tower. The
system can generate saturated or superheated steam. It has been operating
since 2009 to validate the heliostat operation, control system, and receiver
designs.
In 2009 BrightSource continued with the construction of a 29 MWth
enhanced oil recovery facility at Coalinga, between San Francisco and Los
Angeles. It uses 3,822 of the same two-facet mirrors to illuminate a flat
panel steam receiver atop a ∼100 m tower.
In 2009 eSolar brought two demonstration modules of their proposed
10–14 unit commercial plant into operation in California. One utilized a
north and a south facing pair of cavity receivers producing superheated
steam, the other a receiver that consisted of four flat-panel units facing NE,
© Woodhead Publishing Limited, 2012
Central tower concentrating solar power (CSP) systems
255
NW, and SE, SW, with steam superheating taking place near the north and
south corners. The field uses small 1.14 m2 heliostats factory mounted on
truss structures to minimize field installation. In their test/demonstration
facility, they used 24,000 such heliostats, 6,000 each in a north and a south
field for each receiver. They generated superheated steam at over 400°C in
each receiver module, with each such module rated at ∼10 MWth. The objective of these developmental facilities was to test and demonstrate the heliostat concept, control system, and receiver capability. In addition, the two
receiver outputs were ducted to a 5 MWe turbine to demonstrate stability
of the entire system, with the electricity sold to a local utility company
(Meduri et al., 2010). As of 2011, a second-generation demonstration module
of a proposed ten-module 100 MWe molten salt plant is undergoing commissioning in Rajasthan, India.
8.3.2
Commercial power plants
By 2010 there were five or more companies actively engaged in developing
large-scale commercial central receiver projects: Abengoa Solar in Spain
(and a 50 MWe project recently announced in South Africa), eSolar in the
US and in India, Sener-Torresol Energy in Spain, BrightSource (Luz II) in
the US (and Israel), and Solar Reserve in the US. Each had a different
approach to commercialization, involving various heliostat concepts,
receiver designs, working fluids, storage methodology, field configurations,
and plant sizes. Each of these currently has a commercial tower plant in
operation or under construction.
Abengoa Solar, a large Spanish company with numerous trough plants
in operation, was first to proceed to commercial construction and operation
in 2006 with PS10 near Seville, Spain. PS10 is a 10 MWe North (polar) facing
hemi-cylindrical ‘cavity’ receiver at 100 m elevation, generating saturated
steam at 240°C, using 624 canted and focused 120 m2 heliostats. In addition
to powering the turbine, the saturated steam provides 20 MWh of thermal
storage. This was followed by the 20 MWe plant, PS20 (Fig. 8.6), which uses
essentially the same configuration with some improvements in receiver
efficiency, and a one hour storage system. It began commercial operation
in 2009.
eSolar has taken a considerably different approach, using a close-packed
array of co-mounted 1.14 m2 heliostats to irradiate a receiver atop a ‘low’
tower (Fig. 8.7). The commercial concept is for modules to produce 50 MWt
each (10–14 units would produce a total of 500–700 MWth, providing steam
to a 100 MWe turbine with a 50–75% capacity factor respectively, using
molten salt storage). A preliminary design of such a system using molten
salt as the receiver working fluid and for storage at the centrally located
power block is currently (2011) under development for a plant in Rajasthan,
© Woodhead Publishing Limited, 2012
256
Concentrating solar power technology
8.6 The two commercial Abengoa (Spain) saturated steam systems,
PS10 and PS20 (in back), with 120 m2 heliostats in foreground.
8.7 Two 10 MWth modules of the eSolar (USA) Sierra Sun Tower
facility. The receiver in the foreground utilizes four flat panels in an
external square configuration, while the background tower carries a
north-facing and a south-facing cavity receiver. A total of 24,000
heliostats, each 1.14 m2, are factory mounted on support structures
(eSolar, with permission).
© Woodhead Publishing Limited, 2012
Central tower concentrating solar power (CSP) systems
257
8.8 Torresol’s Gemasolar 19.9 MWe commercial molten-salt facility
with heliostats and a local BCS station in the foreground and tower/
receiver in back, probably in preheat configuration. The active receiver
is the central 20% of the white area with the bottom 40% comprising
the BCS target (K. Younglove, with permission).
India, where a second generation demonstration field module is undergoing
commissioning (Pacheco et al., 2011).
Torresol Energy has built a commercial molten salt facility (Gemasolar)
with 15 hours of molten salt storage (Lata et al., 2010), which began
commercial production in the second half of 2011 (Fig. 8.8). The plant
employs 2,650 heliostats of 115 m2, resulting in an external-cylindricalreceiver rating of 154 MWth (delivered to storage), but uses only a
19.9 MWe turbine (to stay within the limitations of the Spanish tariff regulations). In fact, this plant has been operated continuously for several days,
proving baseload capability for a CSP plant. As the local peak-load period
is largely after dark (between the hours 12:00 and 22:00), the storage is
utilized every day to match the peak, making the plant much more valuable to the utility.
The US company, BrightSource, is making good progress (construction
started in October 2010) to satisfy a power purchase agreement for 392 MWe,
comprising three units, at Ivanpah in the California desert near Las Vegas,
taking advantage of a federal loan guarantee. They use 14 m2 heliostats
mounted on posts driven into the ground (to reduce both costs and habitat
degradation) to produce a total of 300–400 MWth per plant on four nominally flat panels facing the cardinal directions. The superheated steam will
drive a conventional turbine or will be directed to heat exchangers to
produce high temperature salt for storage in a two-tank system. The first
component is scheduled for operation in early 2013. 22 months after ground
breaking, in August of 2012, the received was installed on the tower and
over 90% of the heliostats installed on the first unit (Fig. 8.9). The receiver
is being installed on the second unit, as is the heliostat field, at 500 heliostats
© Woodhead Publishing Limited, 2012
258
Concentrating solar power technology
8.9 Phase 1 of 3: BrightSource Ivanpah facility near Las Vegas,
January 2012. 121 m steel tower is ready for receiver installation. Over
20% of the 40,000 heliostat pedestals are installed for this 110 MWe
plant and heliostat installation proceeds at 100+ per day (Michel
Izygon, with permission).
per day. The tower of the third plant is up and field work is underway. In
total, BrightSource has 1.3 GWe of PPAs in place with Southern California
Edison and 1.3 GWe with PG&E. Some of the these plants will incorporate
molten salt storage (BrightSource press release, November 28, 2011 and
August 12, 2012).
SolarReserve in the US licensed the proprietary design information that
Rocketdyne (now Pratt & Whitney Rocketdyne, a subsidiary of United
Technologies Corporation) had developed in the construction of the molten
salt receiver, pumps, thermal storage units, and steam generators of Solar
Two. They are currently (mid 2012) in the initial construction phase (all
permits received and a federal loan guarantee in hand and the central tower
completed, awaiting a receiver) for the first of three 110 MWe (565 MWth)
molten salt plants with an approximately 50% capacity factor to be installed
near Tonopah, Nevada (Fig. 8.10).
Table 8.2 shows a summary of central receiver commercial power
plants.
© Woodhead Publishing Limited, 2012
Central tower concentrating solar power (CSP) systems
259
8.10 SolarReserve demonstration heliostat on Crescent Dunes site,
November 17, 2011, 170 m concrete tower completed February 9,
2012, awaiting addition of cylindrical receiver (SolarReserve, with
permission).
8.4
Design and optimization of central
receiver systems
8.4.1
Determination of system configuration
Before one can initiate the serious design of a CR system, it is essential that
the application be selected, thus determining the power level and the
receiver operating temperature and conditions. For example, a 1 MWth solar
furnace producing a peak flux density of 10 MW/m2, a 10 MWe peaking
plant, a 200 MWe ‘baseload’ plant powering a supercritical turbine, or a
plant to provide 10 MW of process heat at 300°C to a food processing plant
or at 1,000°C to a metal refinery. The details of the application will define
the operating temperature, the receiver working fluid, the load, and the
solar multiple required to satisfy the load by storing energy during available
sun hours. The receiver working fluid may be used directly by the load, used
to charge storage, or used in a heat exchanger, etc. Once these details are
worked out and the design-point power level, output temperature, and
maximum allowable flux density of the receiver defined, serious design of
the CR system can be initiated. Site elevation and latitude are also essential
in order to define the nature of the insolation and the longitude (site) to
define the insolation profile. Of course, an experienced solar engineer
should be involved in all these decisions, or unrealizable conditions may
(will) be defined.
© Woodhead Publishing Limited, 2012
© Woodhead Publishing Limited, 2012
Saturated
steam
Molten
salt
∼1 hour
2009
50 min @ 50%
2006
2010
None
2011
Based on data from developers’ presentations, websites and data sheets.
* Supplemented by estimates by the author: January 2012.
Storage
capacity
First operation
Heat transfer
fluid
278 MWhe
100 m
Flat panel
100 m
140 m
Hemi-cylindrical HemiCavity
cylindrical
(4-panel)
cavity
Saturated
Saturated
steam, 250°C
steam
@ 40 bar
65 m
Dual cavity
external
4-panel
Superheated
steam
185 ha
130 m
External
cylinder
Field area ha
Target height
Receiver
∼
53,500 m2
*∼800
304,750 m2
∼900
2 × 13,600 m2
*∼800
*∼150,000 m2
*∼800
75,216 m2
Irradiance W/m2
Reflector area
2010
None
29 MWt
*220 MWt
2 × 10 MWt
*110 MWt
55 MWt
Thermal power
Process
heat
Chevron
(USA)
19.9 MWe
Gemasolar
(Spain)
5 MWe
pilot plant
Sierra Sun
Tower (USA)
2013
130 m
External
4-panels
N, S, E, W
Superheated
and
reheated
steam
None
3 × 3,000
MWt
∼
3 × 900,000
m2
3 × 130 MWe
Ivanpah
(USA)
BrightSource Industry
20 MWe
PS20 (Spain)
Torresol
10 MWe
PS10 (Spain)
Plant
designation
(plant country)
eSolar
Net electric
power
Abengoa SA
Responsible
organization
Table 8.2 Summary of Central Receiver commercial power plants (2012)
2013
(projected)
3,000 MWht
Molten salt
540 ha
180 m
External
cylinder
950
1,081,000 m2
110 MWe
500 GWhe/
yr
565
Crescent
Dunes (USA)
SolarReserve
2012+
3,500 MWht
Molten salt
*∼100 m
External
4-panels
*10 ×
68,000 m2
100 MWe
prelim.
des
10 ×
50 MWt
Bikaner
(India)
eSolar
Central tower concentrating solar power (CSP) systems
261
Commercialization of CR systems has to some extent been delayed by
the significant capital cost of the very large capacity systems that many
commercial proponents believe to be more cost effective than small or
modular systems. Recognition of this issue has recently led to the design
and construction of smaller systems designed to benefit from a lower capital
cost requirement, as well as smaller and hence easier to handle components,
more rapid progress down the learning/experience curve, easier access to
capital, shorter construction time, and possible use of available off-the-shelf
components (eSolar, 2011). It remains to be seen if these benefits can
outweigh the economies of scale of larger systems. Recently, successful
operation of a one MWe unmanned trough-thermal-electric system showed
promise for small CSP systems in general (Arizona Power Systems, 2006).
Unmanned operation is an important capability in allowing smaller systems
to be economical. Such small systems can be clustered to meet larger applications and also to achieve lower operating and maintenance (O&M) costs.
However, energy transport costs to the central site must be considered
carefully as fewer larger units may be more economical once the energy
transport is accounted for.
8.4.2
The objective function for optimization
The designer of a CR plant must have in mind some criteria, which will
drive the design. In most cases this will result in a system trade-off or optimization process. Because the collectors are expensive, one might select
maximum plant efficiency, achieved via high output temperature to allow
better Carnot efficiency, but this is likely to call for more accurate (possibly
more expensive) heliostats and is sure to result in higher receiver thermal
losses and optical spillage and more expensive receiver materials, and generally a more expensive power block, though higher efficiency can reduce
costs of auxiliaries such as cooling. Nonetheless, some cycles such as the
supercritical CO2 Brayton turbine promise very high efficiency at quite
achievable temperatures.
Aiming for lowest capital cost, on the other hand, may result in the choice
of less accurate heliostats, thus a larger receiver to overcome spillage, higher
thermal losses or a lower operating temperature, and a poor system performance, perhaps resulting in a higher levelized electricity cost (Pitman and
Vant-Hull, 1986). One could also choose to design the plant for a particular
day or time, resulting in relatively poor performance at other times and on
an annual basis.
For a commercial CR plant, we assume the customer for the thermal
energy produced is a power plant or a process heat application. The exact
application will then define the specific required power and capacity factor,
and a likely temperature range. The solar plant is then designed to meet
© Woodhead Publishing Limited, 2012
262
Concentrating solar power technology
this annual demand at the lowest cost per unit of energy. The appropriate
objective function is then a cost benefit ratio: total capital cost of the
installed plant plus present value of operations and maintenance over the
projected life of the plant all divided by the net output of the plant over its
design lifetime: i.e., the lifetime cost per MWhr of the thermal energy delivered (Lipps and Vant-Hull, 1978). It is important to note that multi-year
average diurnal insolation should be used in the optimization. A single
selected year is unlikely to be representative and, while a ‘typical year’ may
be constructed, an insolation model based on long-term atmospheric properties is easier to construct and understand.
An alternative objective function that is frequently used (Dellin and Fish,
1979) is levelized cost of electricity (LCOE) (discussed in detail in Chapter
2), defined as the unit cost at which the product must be sold to provide a
defined rate of return over the life of the plant. This is a more complex
analysis and one must define the financial details completely and into the
future (and they strongly affect the result). Further, the cost and performance details of the non-solar balance of the plant are required. Hardly
any of these issues are solar related, and while LCOE is a quantity easily
recognized by utilities and banks, financial quantities vary markedly over
time, and the performance details of the subsystems may be difficult to
obtain and implement in a meaningful manner. Nonetheless, the emergence
of dispatchability as a primary advantage of CSP means a financial analysis
method that can model the benefits of time of day dispatch (whereby a kWh
of electricity may be worth more in the afternoon/evening than in the
middle of the day, for example) will be of most value for investors.
SANDIA has developed codes such as Solergy (Stoddard et al., 1987) to
accept the diurnal efficiency curves for the optimized fields produced by
programs such as RCELL or DELSOL, and do a detailed calculation of
the energy or electricity produced during a representative or average year.
Solergy has incorporated the thermodynamic quantities and the financial
parameters required to develop an LCOE. A good summary of software
and codes for analysis of concentrating solar power plants is provided in
Ho (2008), though even more models have emerged since then.
The effects of inflation must be accommodated. As various subsystems
or components may be costed at different times, it is reasonable to multiply
each cost by an appropriate inflation factor. The familiar consumer price
index offers a generic method for accommodating inflation; however, at the
component level a more appropriate method may be to use an engineering
source such as the Chemical Engineering Plant Cost Index (CEPCI, published monthly by the American Society of Chemical Engineers). Costs
generated at various times can be compared and/or used coherently in a
study if each cost is divided by the index value appropriate for the date it
is generated, and multiplied by the current index value. This allows ‘legacy’
© Woodhead Publishing Limited, 2012
Central tower concentrating solar power (CSP) systems
263
costs from earlier work to be used alongside current costs. Many countries
have such resources. A more complex consideration can be the incorporation of international currency exchange rates.
8.4.3
Items to include in the cost function
The numerator in the objective function is money invested over the life of
the plant. This may be considered in three categories: preliminaries and
fixed, purchase and construction, and finally, operations and maintenance
(O&M).
Fixed costs such as permitting, design, and access
These should include most general costs such as plant design; environmental and political permitting; central facilities such as control room, fire and
safety facilities, maintenance shop; and access roads and transmission lines
to the site – which may be substantial if the site is isolated. As each of these
quantities can cost over a million dollars, they must be considered carefully
and should be prorated to the various subsystems of the facility. In fact, they
are a primary reason energy from small, stand-alone plants tends to be more
expensive. This issue may be somewhat mitigated if a number of small
plants are integrated or co-located with one another or with the energy
user.
Capital costs
The major solar-related capital costs of a CR plant are the heliostats, the
receiver, the tower, the thermal transport system, and the thermal storage
system if included. It is important that each of these be not just the purchase
price of the item, but include the installation cost (transportation, foundations/support structures, construction or installation, connection to power
and control system, quality control, and testing). The cost of a heliostat on
the factory floor may be little more than half of its final cost as an installed,
wired, aligned, tested, calibrated, and operational heliostat at a distant site.
Similarly, the vertical piping cost must include transportation to the site,
hangers, an allowance for expansion bends, on-site welding and inspection,
insulation, heat trace if needed, temperature sensors, valves, etc. It is often
a surprising but important fact that the actual cost of the constructed active
receiver (tubes and headers, etc.) is only a small part of the cost of the
receiver. The majority of the cost is in the strong backs and hangers the
panels are mounted on, the support structure, insulation, temperature
sensors, valves and controls, header shields and ovens (if needed), and the
transition from the tower to the support structure. These costs are even
© Woodhead Publishing Limited, 2012
264
Concentrating solar power technology
higher for a cavity receiver, which must also maintain the receiver tubes
and headers in a high temperature enclosure.
To allow for optimization of the CR plant, the cost of each subsystem
must be provided at the outset in terms of the appropriate design variables
in simple relations valid over the range of practical values. For example,
towers of a few heights based on designs appropriate for the local earthquake regime and assumed top loading due to expected receiver size,
weight, and wind speeds can be developed. Then the tower cost for any
height within an appropriate range can be approximated with a power law
or exponential fit, after any obvious fixed or linear costs are accounted for
(soil testing, safety lighting, elevators, etc.). Of course, the final design will
provide a more detailed and precise tower design and cost.
Land
A good solar site will likely be relatively isolated, have a peak direct-beam
insolation of the order of 1 kW/m2, and an annual beam insolation (DNI
– direct normal insolation) preferably >2,000 kWh/m2/yr, be of suitable size
(including buffer), moderately level (less of a requirement for central
receivers than troughs) and suitable for construction. It must also have
easily resolved environmental constraints, good and guaranteed access to
nearby transmission lines able to accept the power and to roads adequate
to handle construction equipment, and a willing seller. Such ideal land may
sell in the USA for $1–2/m2 (or $4,000–8,000/acre). While much cheaper
land may be available, satisfying or coping with all the conditions above can
be very expensive, and time consuming, while permitting may be impossible.
Gently rolling land can be accommodated, but will significantly complicate
the layout of heliostats. Sloping land, up to perhaps 10°, can be accommodated in the system design of a central receiver plant with varying effects
such as: later sunrises/earlier sunsets due to virtual horizons, shifting the
centroid of the field downhill, or moving the diurnal energy output curve
toward morning hours if the slope is up to the west or to better meet an
afternoon peak load if up to the east.
Heliostats
Heliostats must be accurate enough to not degrade the solar image excessively (beam error less than ∼2–3 mrad) and rigid enough to sustain gravity
and wind loads while maintaining this accuracy. Such heliostats are presently
estimated to cost ∼$200/m2 in 2010 US dollars (Kolb et al., 2007; Kolb, 2011)
and most cost targets are between $100 and $130/m2 once volume production and learning curve effects begin to take effect (Mancini et al., 2011). The
long-term DOE SunShot goal is $75/m2, which will be difficult to achieve.
© Woodhead Publishing Limited, 2012
Central tower concentrating solar power (CSP) systems
265
The optimum heliostat area is hard to define as it involves a trade-off
between many effects such as reflector support deflections under gravity
and wind load, spillage (beam size at the receiver), and number of control
systems, etc., to be built and maintained. Current ‘commercial’ heliostats
range from 1 m2 to 150 m2; however, designs up to 200 m2 have been developed, and some companies have considered ganged or autonomous heliostats much smaller than 1 m2. See Chapter 17 for a more detailed analysis.
To be cost effective, a heliostat must make a net positive contribution to
the economic performance of the plant. For example, if the land cost associated with a heliostat near the center of the plant is 2% of the heliostat cost,
near the field boundary where the heliostat density is typically one-third that
at the center of the field, the effective land cost will be 6%. Atmospheric
attenuation between the heliostats and the receiver has a similar effect as
does spillage losses. The cost of conventional wiring and control systems also
has a negative impact due to the larger radial spacing far from the receiver.
Thus the position of the outer boundary and heliostat losses and associated
costs need to be traded off to optimize the objective function.
Present value of subsystem operations and maintenance (O&M) costs
Annual operations and maintenance (O&M) costs may amount to a quarter
of the annual charges for the capital costs. As a result, it is important to
include such costs in the optimization. However, it is inappropriate to
simply multiply the overall plant cost by 125%, as O&M cost comprises
both fixed and variable components and varies widely from subsystem to
subsystem. For example, the Utility Study of the 1980s (Utility Study, 1988)
used the consensus estimates for the annual O&M relative to each subsystem capital cost for that molten salt system: tower 0.2%; receiver 2%; vertical piping 1%; feed pump 5%; heliostats 1%. For the 39 m2 heliostat
considered at that time the costs were split equally between the reflectors
(area) and the tracking system (number). Experience since then will allow
more operational values to be developed for each subsystem, but the argument is unchanged. Note that for comparison to capital costs, these percentages must be multiplied by a present value factor which may range from 10
to 20 times the annual costs depending upon the real rate of return assumed
(see Chapter 2).
8.4.4
Choice of performance criterion
Design point or annual
The objective function for optimum design of a CSP plant has costs in the
numerator and performance in the denominator. There are a number of
options for definition of the performance, and some are much better than
© Woodhead Publishing Limited, 2012
266
Concentrating solar power technology
others. Choosing performance at the design point will lead to a lower cost
system with excellent performance at that time, but relatively poor performance at other times throughout the year, and consequently a lower annual
energy output. While this may be suitable for a solar furnace with short
seasonal operating times, it is inappropriate for a power plant. Here the
requirement is lowest LCOE, so the denominator must be delivered annual
energy.
Incident, absorbed, or delivered energy
The energy incident at the receiver is reduced by the receiver interception
and reflectivity (1 − absorptivity); radiation and convection losses (although
conduction losses are negligible); piping thermal losses; and the parasitic
losses of the feed pumps, valve and heliostat actuators, etc., required for
operation of the solar complex (converted from electrical to thermal by
multiplying by the heat rate of the turbine or charged directly as purchased
electricity). The result is the net delivered thermal energy available to the
turbine or to charge the thermal storage unit.
Inclusion/effect of time-of-day pricing, sloped fields
If time-of-day pricing is in effect, it can be used by the plant designers to
improve the annual revenue of the plant by configuring the field layout so
it is more effective during the period of peak electrical price. This can be
used with or without storage. A more effective approach is to value the
sunlight proportionately to the time-of-day price prior to optimizing the
field. Thus, high value sunlight will play an enhanced role in establishing
the heliostat spacings everywhere, and in defining the boundary of the plant.
With an afternoon peak price, this will lead to an east-biased field both in
heliostat density and field boundary. A field that slopes up to the east will
have a similar effect and is favoured in case of afternoon peak pricing. An
upslope to the north (or south in the southern hemisphere) will produce a
somewhat more effective field, but the cost of installing heliostats on a
sloped field must be accounted for and a more polar bias of the field must
be accommodated.
8.4.5
Effect of constraints on optimization
Any constraint on the system design will have a negative impact on the
resulting optimized field/system, and an even greater impact if the constraint is imposed after the optimization. The unconstrained optimization
process strives for the lowest LCOE. Consequently, for a specific tower
height and receiver aperture, it defines the heliostat spacing and number
(field boundary) and evaluates the design point power and annual output
© Woodhead Publishing Limited, 2012
Central tower concentrating solar power (CSP) systems
267
power. One can achieve a lower design point power and perhaps higher
efficiency by trimming heliostats from the boundary of the field, but the
LCOE will be increased as cost-effective heliostats are removed. Alternatively, one could constrain the design point power at the onset and the
optimization process will lead to a lower tower, smaller receiver, and intermediate area field – at an intermediate LCOE. Clearly, the least constrained
system will be most cost effective, and the best time to introduce any constraint is prior to the optimization. Definition of a polar-facing aperture is
clearly a severe constraint, but the use of two or three cavities to enhance
the azimuthal acceptance angle can ameliorate the constraint somewhat.
8.5
Heliostat factors
8.5.1
Beam errors
An ideally focused optical mirror operating on-axis can form an image of
the sun on the receiver at a distance of one focal length. As the sun is a
distant object with an angular diameter of 9.306 mrad, the image will not
be a point, but have a diameter of 0.0093 times the focal length. At 1,000 m,
this amounts to a 9.3 m diameter, and would display the limb darkening of
the sun (due to scattering in the solar atmosphere) and the solar aureole
(light appearing outside the limb of the sun due to scattering in the Earth’s
atmosphere). For the purposes of CSP systems involving concentrator
optics, we are interested, not in forming an image, but in collecting most of
the energy cost effectively. Thus, the sun may be conveniently represented
by a Gaussian function which is a least square fit to the limb-darkened solar
image, resulting in a second moment (sigma) of 2.770 mrad, but will overestimate the peak flux density by about 16%. The resulting Gaussian ‘spot’
at 1,000 m will have a diameter of 5.54 m at the one-sigma radius (relative
intensity = 0.607) and a diameter of 11.08 m at the two-sigma radius (relative intensity =0.135). The energy beyond this diameter (i.e., the spillage for
a circular target) would be mathematically equal to the relative intensity at
the defined radius, assuming the Gaussian is a reasonable approximation to
the solar aureole. A better fit to the true sunshape can be accomplished by
adding to the second moment (sigma2) higher moment terms such as fourth
moment (a measure of kurtosis), sixth moment, etc., in a Hermitian series
(Walzel et al., 1977), but if beam errors are comparable to the second
moment of the sunshape the improvement is marginal and the complication
significant. For detailed flux analysis, it is useful to include terms in the
Hermite series to sixth order.
Beam errors encompass all imperfections in the heliostat system and are
expressed in terms of the divergence half angle of the beam leaving the
heliostat if illuminated by a true point source. It is assumed that all systematic errors are identified and corrected by normal O&M procedures so the
© Woodhead Publishing Limited, 2012
268
Concentrating solar power technology
remaining errors are essentially random, at least if averaged over an array
of heliostats. Residual mechanical errors, which may affect the orientation
of the mirror, are doubled upon reflection. Such errors include tracking
errors, effects of wind or gravity on the heliostat, etc. It is very convenient
to handle all of these random errors by summing them in quadrature (after
doubling mechanical errors) to obtain the ‘total beam error’, and then
adding them in quadrature to the second moment of the limb-darkened sun
(2.770 mrad) to form a ‘degraded sun’. This avoids all the issues of convolving a multitude of errors on a multitude of heliostats. This also provides a
useful reference for judging the importance of beam errors, if the total beam
error is equal to that of the sun (2.770 mrad) it will produce a degraded sun
with a sigma of 1.41*2.770 = 3.92 mrad, if the beam error is much less, it
will be relatively unimportant, if much larger it will be dominant. The idea
of a degraded sun gains validity from the fact that we usually have thousands of heliostats, so individual instantaneous displacements of individual
beams or facets can be treated as random variables. It does give up the
ability to predict realistic images from a single heliostat. Higher order
moments of the sun and of the errors can also be incorporated via the
Hermite method mentioned above.
8.5.2
Heliostat size
The discussion of sun-shape and beam errors provides the optical rational
for considering heliostat size. An optically flat heliostat will simply project
an image of the sun from each point, forming an image of the mirror with
the effects of umbra and penumbra representative of the solar beam. So long
as the angular size of the mirror as seen from the receiver is small compared
to the angular size of the sun, the intensity of the image will not be reduced
significantly compared to a focused mirror. If the mirror is comparable to or
larger than the projected solar image, there will be a serious decrease in
intensity from a flat mirror compared to a focused mirror, accompanied by
a corresponding increase in the spot size as discussed in Section 8.5.3.
8.5.3
Focusing and facet canting
The spot degradation caused by large, flat mirrors can be largely recovered
by cutting the mirror into facets (which also makes them easier to handle)
and canting each one to superimpose their images at the receiver (which
carries some cost in hardware and labor). Alternatively, the mirror, or the
individual facets, can be curved to focus their beam near the receiver
surface (often termed as ‘to slant range’) in order to reduce the effect of
the facet diameter on the solar spot. As it is much easier to bend glass into
a cylinder than a sphere (due to Poisson ratio effects), frequently mirrors
© Woodhead Publishing Limited, 2012
Central tower concentrating solar power (CSP) systems
269
are focused only on the long axis. Whatever the optical geometry (figure),
the image (from a point sun) of the error-free mirror can be projected to
the receiver via ray-trace and the second moment (and higher moments)
of this ‘spot’ can be calculated and added in quadrature to that of the
degraded sun (Walzel et al., 1977), or the image may be generated by
detailed ray tracing.
8.5.4
Off-axis aberration
Either canting or focusing results in an effect known as off-axis aberration
(Rabl, 1985, p. 177). Effectively the focal point is modified by the angle of
incidence (ι) of the sunlight on the mirror, so in the transverse direction
the focal length (f) is reduced by cos ι and in the sagittal direction increased
by 1/cos ι, forming two line foci with a circle of least confusion between
them near f. For a mirror of diameter w and incident angle i, the diameter
of this circle is w(1 − cos i). At an off-axis ι of 60°, this leads to a heliostat
image from a ‘perfectly focused on axis to slant range’ parabolic (or small
spherical) mirror of one-half the mirror dimension (to be added in quadrature to the sun image and the effects of beam errors discussed above). The
situation becomes more complicated as one departs from the nominal focal
length.
8.5.5
Effects of tracking mode
There are a number of possible mounting systems to accomplish the tracking of the heliostat image onto the receiver. Most common is the elevation–azimuth (el-az) system, in which the assembly is rotated in azimuth,
and above that is mounted the orthogonal elevation axis carrying the mirror.
While simple and economical to build, this scheme has several disadvantages. One is that the reflector rotates in azimuth but does not retain its
orientation with respect to the sun-mirror-receiver plane. Consequently,
the angle-of-incidence effects mentioned above cannot be resolved, so the
(nearly flat) mirror is generally given the optical figure of a sphere as the
best option. Alternatively, the mirror can be given a third axis of rotation
(about its center) and rotated continuously to retain the correct orientation
with respect to the sun-mirror-receiver plane (Zaibel et al., 1995; Chen
et al., 2001). A tracker with the first axis pointing toward the receiver, or
spinning-elevation tracking system accomplishes this also (Zaibel et al.,
1995). The mirror can then be given an appropriate off-axis parabolic optical
figure, obviating the cos ι effect.
In any case, to avoid collisions with its neighbours in case tracking control
fails, el-az mirrors require a ‘clear-out circle’ with diameter equal to the
maximum dimension of the horizontal mirror (D for a circle, 1.41 S for a
© Woodhead Publishing Limited, 2012
270
Concentrating solar power technology
square, (L2 + W2)0.5 for a rectangle, etc.). Typically 30 cm or so (for a large
heliostat) is added as a safety factor, or more if the axis of rotation does
not pass through the center of the horizontal mirror. This ‘mechanical limit’
may increase the required separation of heliostats, a problem particularly
in the most effective portion of the field where heliostats tend to be most
crowded. Conceptually, collisions of more closely spaced heliostats can be
avoided by computer algorithm, but failures will be costly.
An elevation/polar-axis mounting system carries a mirror which does not
rotate in azimuth. This allows close packing of the mirrors giving twice the
maximum mirror density for square mirrors compared to the el-az mount.
It also tracks at a constant rate about the polar axis. The orthogonal axis
requires only slow corrections for seasonal changes in solar declination.
A mounting system with the first axis pointing toward the receiver can
rotate at a constant rate, so a second axis is always perpendicular to the sun
vector (Francia, 1968). The mirror in this case will always retain the correct
orientation with respect to the sun-mirror-receiver plane. Thus, if given the
optical figure of an appropriate off-axis parabola, there will be no cos ι
effect.
It is important to consider the added cost compared to the advantages of
these, and other, exotic mounting systems before selecting them.
8.5.6
Effects of heliostat size on heliostat cost and
other factors
All of these effects must be taken into account when selecting the heliostat
size. Clearly, heliostat dimensions affect the concentration and spillage. In
addition, larger heliostats benefit from economy of scale (to a point) while
smaller heliostats will require a greater number of pedestals, controls, actuators, etc., and so benefit more from learning/experience curve effects. Heliostats from 1 to 200 m2 have been developed, with no consensus as yet (see
Chapter 17). Clearly small systems will benefit from smaller heliostats, but
O&M costs must also be considered.
8.5.7
Reflectivity and cleanliness
Of course, all heliostat mirrors suffer losses due to imperfect reflectivity. This
can vary from 3% for a well protected front surface silvered mirror to 4–5%
for a second-surface mirror employing water-white glass, to 10–15% for
thicker, high-iron glass. This pristine reflectivity will remain quite stable in
time for any acceptable mirror, but may well be degraded by 1% or so by
surface dents or scratches, etc., by the time it is installed in the field. Any
decrease in reflectivity degrades the performance of the heliostat, so the
added cost/m2 of better reflectivity should be compared, not to the cost/m2 of
the mirror, but to the cost/m2 of the heliostat field (typically five times larger).
© Woodhead Publishing Limited, 2012
Central tower concentrating solar power (CSP) systems
271
Once installed, the reflectivity of corrosion free mirror declines by perhaps
2% per month, partially due to easily removed dust and partially due to a
harder to remove surface film (depending on the site and atmospheric conditions). Cleaning equipment is costly and somewhat labor intensive, so one
must trade the washing frequency (cost) against the cost of lost reflectivity,
as zero loss implies infinite cost. Rinsing to remove the dust can be done with
a boom truck and is quite fast, while scrubbing to remove the surface film
requires a more costly brushing rig and is slower. If the mirrors are scrubbed
at the optimum rate, the loss in average reflectivity may be limited to about
6%, while if a lower-cost rinse cycle is added to simply remove the dust
between scrubbings, the loss in average reflectivity with optimum cleaning
may fall to 3.5% (depending on the environment, rig costs, and labor costs).
By trading decreased washing costs against the cost of the added heliostats
required for the dirty field to provide the same energy, an optimum average
reflective loss can be determined. The optimized average reflectivity should
be used in determining the required size of the heliostat field, or production
goals will not be met (Kattky and Vant-Hull, 2012).
8.6
Receiver considerations
8.6.1
Cavity vs flat vs cylindrical receivers
Field constraint
A typical design constraint is to define a restricted receiver aperture, e.g.,
a billboard or a cavity receiver rather than a cylindrical receiver. While a
cylinder has a field of view of 360°, a billboard is restricted to 180°, and
often much less, although inclining the aperture forward can put a larger
ground area within the field of view. A polar-facing billboard with a 120°
field of heliostats will perform very well near noon, when most of the
heliostats experience near 0° angle of incidence of the sunlight and so have
a ‘cosine effect’ near unity (a combination of the foreshortening of the
heliostat by cos ι and off-axis aberrations). However, as the solar azimuth
changes (at other times of the day) the performance falls rapidly. In contrast, a system utilizing a cylindrical receiver will have moderate performance near noon (due to the reduced cosine experienced by heliostats on
the equatorial side of the receiver). However, as the azimuth of the sun
moves away from its noontime value (by more than 100° in the summer),
the nearly circular nature of the surround field tends to maintain the
average value of the cosine over the field, so the annual performance for
the surround field exceeds that for the polar field of the billboard or cavity.
For small systems, a billboard may be significantly cheaper than a cylindrical receiver, and for high temperature working fluids with low heat transfer
coefficients (gases), a cavity may be required to reduce thermal losses, but
© Woodhead Publishing Limited, 2012
272
Concentrating solar power technology
the effects of the resulting constraint on the field performance must be
taken into account.
Reflective, radiative, and thermal loss of the cavity
It is not true that the aperture of a cavity is a ‘black hole’. In fact, a positive
feature of cavities is that the sunlight entering the aperture may be distributed over the absorbing surface to reduce the issues of excessive flux density.
This process is usually accomplished by having much of the sunlight reflected
or scattered from adiabatic surfaces. This results in a significant amount of
diffuse solar light bouncing around in the cavity, some of which finds the
aperture and is lost, effectively reflected. A secondary effect is that much
of the interior of the cavity absorbs and reradiates light at a temperature
above that of the heat transfer surface, and a significant portion of this
infrared energy is also emitted from the aperture. Finally, the entire interior
of the cavity is in contact with the enclosed air, heating it to a very high
temperature and instigating turbulent convective circulation rolls. Depending on the orientation of the cavity, such circulation can carry heat to the
aperture and out, strongly encouraged by ambient wind. The result is that
a poorly designed cavity receiver can experience greater losses than a billboard or cylindrical external receiver. For small receivers a window can
help, but this has difficulty with scaling up for commercial scale systems.
Cost and weight
Cavity receivers have specific application for very high temperature systems
where there is benefit in having a small ratio of aperture to absorber area
and where high fluxes can be absorbed. However, a cavity must both support
and enclose the heat-collecting surface, it will clearly be larger than the
support structure required for an external receiver where total absorbed
flux is the same. In addition, it must be well insulated to reduce conductive
losses through its large surface area. The result is a larger, heavier, and more
expensive housing for the actual receiver. Cavity receivers are usually
designed for smaller capacity systems, because the limited cone angle
demands a higher tower. It is routinely found that the receiver tubing,
headers, valves, etc., represent only a small fraction of the cost of the entire
assembly, even for an external receiver.
8.6.2
Effect of allowable flux density on design
It is quite easy to design a heliostat field which produces a peak concentration of several thousand suns (several MW/m2). However, most commercial
system receiver designs can only tolerate a flux density in the order of
© Woodhead Publishing Limited, 2012
Central tower concentrating solar power (CSP) systems
273
1.0 MW/m2 (e.g. for molten salt) due to the heat transfer characteristics of
the fluid. Thus, a larger receiver must be used than the size of the minimum
achievable focal spot. The aspect ratio of an external receiver (height/
diameter) should be greater than one to allow close-in heliostats, which see
a severely foreshortened receiver, to be effective. In addition, it is typical
that the receiver designer requires a significant tube length to achieve
design temperature at the high heat-transfer-media flow rates required to
enhance the allowable flux density. The final receiver aspect ratio will typically be between one and two. The allowable flux is then achieved by defining several (2–5) aim levels to each of which specific fractions of the
heliostats are assigned. A smart strategy that estimates the beam radius, and
moves the image associated with heliostats assigned to the outer aim levels
so their beam just remains on the receiver, will introduce very little spillage
while reducing the peak flux by 1.5–4.0 times.
A second requirement may be that the ratio of north to south side flux
be constrained to maintain some balance on the receiver panels, and to
reduce transverse gradients on each panel. This can be accomplished by
enhancing the field boundary on the ‘weak’ side at the expense of the other.
A more sophisticated approach (Vant-Hull and Pitman, 1990) involves
modifying the objective function of the optimizer with a function (for
example, cosine) of the azimuth angle to ‘inform’ the field of the new constraint. At some cost, this penalizes the polar heliostats and benefits the
equatorial heliostats so the density and extent of the fields are modified,
retaining the concept of an optimized field but relieving the receiver, saving
money. The flux density gradient across each panel may also be a concern,
but this is largely defined by the ratio of anti-sun to sun-side flux which can
be modified as above. Transverse shifts of smaller images can reduce the
local transverse gradient somewhat.
8.6.3
Emissivity vs absorptivity vs temperature
At the reduced temperatures and concentrations typical of linear systems,
reduced emissivity absorbers (solar selective surfaces) can be used effectively to limit the radiative losses. However, it is difficult to design selective
absorbers that will survive when exposed to the atmosphere while operating
at temperatures over 500°C. As α = ε at each wavelength, and both the sun
and a hot receiver emit significant energy in the near infrared range, reducing the emissivity here will also reduce the absorptivity. A sharp change
from α ≈ 1 to α = ε << 1 between wavelengths of 1.5–2 microns is required.
It is most important to keep in mind that the objective is to improve the
net energy captured in the receiver, and at a solar flux density of 1 MW/m2,
a reduction in α of 0.01 (or 0.02) loses 10 (or 20) kW/m2 requiring, at a
surface temperature of 527°C (800 K), a reduction in ε from 1.0 to 0.56 (or
© Woodhead Publishing Limited, 2012
274
Concentrating solar power technology
0.12) to achieve (1 − ε)σT4 > 10 (or 20) kW/m2 just to break even. As higher
temperature receivers are anticipated (e.g. to drive a supercritical turbine)
and the surface temperature significantly exceeds the working fluid temperature, thermal losses will increase, but so will the overlap of the solar
and IR spectra, making the task even more challenging.
8.7
Variants on the basic central receiver system
8.7.1
Polar vs surround fields
A billboard or cavity receiver naturally results in a polar field (generally
termed a north field in the northern hemisphere), i.e., one with heliostats
on the polar side of the anti-polar facing receiver, due to much smaller
incidence angles on the antisun side of the receiver. Such a field will have
very good performance at noon but will fall off rapidly at larger solar azimuths due to the increased values of incidence angle (decreased cosine)
on nearly all the heliostats. In contrast, a cylindrical receiver will be best
matched by a surround field, usually biased toward the pole to take advantage of the better cosine effect, having a pole/equator energy ratio between
1.5 and 2. This system will exhibit a lower average performance at midday
compared to the polar field, but at large solar azimuth, the east or west
fields will contribute strongly to the energy collection, as will the weaker
equatorial field. The net result is that, for a given design point power, annual
energy collected is substantially higher for the surround field with a cylindrical receiver, and thus it is superior on a cost-per-unit-of-energy basis.
8.7.2
Beam-down systems
In a beam-down CRS (Fig. 8.1(d)), a secondary mirror is inserted between
the heliostat field and its elevated focal point to redirect the collected sunlight to the ground. In order to produce a focused spot near the ground, the
mirror must be a hyperbola (or if above the primary focal point an ellipse),
and must be large enough to intercept the entire cone of light from the
field. Such a system is used effectively in astronomy and is known as a
Cassegranian system.
The first response to a tower or even a dish system is often: ‘If you have
the receiver supported on a tall tower and the engine down on the ground,
why not use a Cassegranian system with a small elevated mirror and the
receiver down near the ground where it is easy to connect to the engine and
to service? It works in astronomical telescopes, why not here? And it would
save much of the heat transfer piping.’ The answer is that astronomical telescopes universally have a small f number (ratio of aperture diameter to focal
length), while solar collectors often (usually) have an aperture diameter
several times the focal length. This is because stars really are unresolvable
© Woodhead Publishing Limited, 2012
Central tower concentrating solar power (CSP) systems
275
point objects so magnification of the image is really a virtue, not a problem.
In contrast, the sun has a diameter of 9.3 mrad, and we are trying to form a
concentrated ‘image’. In a Cassegranian system, the elevated ‘hyperbolic’
mirror forms an image of the virtual object (the image which would be
formed by the converging light from all the heliostats at the focal point of
the array). The real image formed (perhaps near the ground) from this
virtual object is magnified by the ratio of the distance from the secondary
mirror to this real image divided by the smaller distance from the virtual
object to the secondary mirror. The convex-downward hyperbolic Cassegranian mirror must be large enough to intercept the entire cone of rays
directed toward the focal point, so in order to have a small secondary mirror
(relative to the area of the collector field), the secondary mirror must be near
the (virtual) focal point of the array and far from the ground (actually, the
receiver aperture which may be elevated somewhat). However, the ratio of
the distance from the virtual focal point to the vertex of the secondary and
the distance from the secondary to the receiver aperture (which may have a
small elevation) is what defines the linear magnification of the assembly,
which would be 10 to produce a secondary mirror area 1% of the field area
(3 or 4% of the collector area due to heliostat spacing). The resulting area
magnification of the virtual object 10 × 10 = 100 reduces the flux concentration by 100 (from a few thousand to a few tens), and is generally disastrous
to the objective of obtaining a high temperature with low loss. Moving the
secondary mirror downward improves the final concentration, but requires
an impractically large mirror and three or four support towers. Moving it to
the halfway point recovers the full concentration at the ground (i.e., the
required flat mirror images the virtual image exactly). However, then the
secondary must be half the diameter of the field, or have an area essentially
equal to the area of the mirrors (assuming an average ground coverage of
25%). It will also produce a large shadow on the field. Supporting the
receiver above the ground level reduces the magnification, and so should be
considered if the beam-down configuration is required.
8.7.3
Use of compound parabolic concentrators
Some of the concentration lost by the beam-down configuration can be
regained by use of compound parabolic concentrators (CPCs) at the
receiver aperture (Rabl, 1985, p. 191). The acceptance angle of the CPC
must be set to accept the entire beam reflected by the secondary, i.e., at the
cone half angle given by secondary radius/secondary distance (from the
CPC aperture). This angle, φ, also defines the concentration of the 2D
(conical) CPC as (1/sin φ)2. If the final image from the secondary is large, a
‘flies eye’ array of 7 or 19 CPCs may be required, but each should view the
entire secondary to avoid excessive rejection of light outside the view angle.
While workable designs can be developed, the added two reflections, the
© Woodhead Publishing Limited, 2012
276
Concentrating solar power technology
extra expense and complication of the CPC array, and the cost of the secondary and of supporting it on multiple (although smaller) towers nearly
always result in a non-competitive design (Vant-Hull, 1991).
The use of a CPC with a normal central receiver system is questionable
due to the large angular extent of the field (as seen from the tower-mounted
receiver.). However, at the Sandia 5 MWth test facility, several linear CPCs,
each viewing a different sector of the north field (NE, N, and NW), were
stacked to enhance the flux on a long, narrow panel receiver under test. This
imaginative solution worked very well, but it was also quite expensive.
An important characteristic of a CPC is that the concentrated light at the
exit aperture is diffuse light; that is, it expands into the full 2Pi steradians.
This can be ameliorated somewhat by truncating the exit aperture, which
also reduces the concentration.
8.7.4
Optical beam splitting
It is sometimes suggested that a beam splitter be used to divide the energy
at the receiver into two or three wavelength bands, so that each beam could
be used most effectively, especially where the energy conversion device
performance is sensitive to wavelength. A possible motivation for doing this
would be to apply selected wavelengths to a concentrating photovoltaic
conversion process and to use the remainder for thermal conversion. There
are numerous research projects, but thus far no commercial implementation. The simplest application would be to mount transparent solar cells in
front of the receiver, with all light not used by the PV cells transmitted on
to the receiver. While difficult to implement in large systems, this could be
used effectiviely in dish systems. A thermal receiver is generally a wide band
absorber, and would absorb the entire solar beam quite effectively – nothing
is really going to waste. Where a concentrating photovoltaic device produces electricity more cost effectively than a thermal device, then it would
be up to the thermal device to make use of the remaining spectrum, if sufficient energy remains to warrant the cost. In addition, there is little energy
in the UV rays to start with, and most of this is absorbed in the initial reflection (due to impurities in the glass, though not an issue with front surface
reflection) so very little reaches the receiver. Perhaps it would be possible
to separate the IR from the visible beam at the receiver, but it is hard to
imagine a better use for this IR than to contribute to the energy absorbed
by a thermal receiver.
8.8
Field layout and land use
Any straightforward land constraint simply reduces the energy available at
the receiver and increases the cost/benefit ratio of the design. If allowed,
© Woodhead Publishing Limited, 2012
Central tower concentrating solar power (CSP) systems
277
the system will add heliostats at the boundary of the field, increase the
density of heliostats everywhere in the field, and increase the height of the
tower to compensate, but the LCOE will increase.
It is interesting that a well-defined, optimized, unconstrained field (with
a cylindrical receiver), characterized by the annual output/m2 of glass, is
nearly circular, but offset from the tower toward the pole by 10–20%,
depending on latitude. However, if land use is at an absolute premium, a
field boundary that is defined by the output/m2 of ground ends up being
very nearly a circle centered on the tower.
8.8.1
Field layout for optimized systems
It has been found empirically (Lipps and Vant-Hull, 1978) that a circular
radial-stagger (RS) heliostat layout is the best compromise. Far from the
receiver, blocking from those aligned neighbors nearer the tower predominantly controls the radial heliostat spacing. The RS configuration
overcomes this issue by moving the nearest radial neighbor two circles
away. The annual shading footprint is more nearly circular, tending toward
the pole. This leads to a requirement that nearest heliostats be at least
two diameters apart in azimuth nearly everywhere. This also allows the
distant heliostats to avoid blocking by ‘peaking through’ between the
neighbors in the first inner circle. Near the tower the heliostats tend to
operate close to horizontally to illuminate the receiver. Consequently, there
is essentially no blocking and little shading, and they can be much more
closely spaced. Here the issue of mechanical limits must be imposed to
prevent errant heliostats from hitting one another. In this region, round
heliostats have an advantage, as do polar mounting systems, which never
allow the mirror to move outside its initial footprint. Near the tower, the
RS layout becomes inefficient because of the rapid variation of azimuthal
separation imposed by the expanding circles. One alternative is to use a
hexagonal close-packed configuration near the tower, converting to the RS
configuration at a radius of one or two tower heights and merging the
interface smoothly (Walzel, 1978). A second alternative is to simply pack
the heliostats closely on circles surrounding the tower and space the circles
as closely as possible to allow servicing and obey mechanical limits, as was
done at Gemasolar. At some radius, blocking will become significant and
the RS configuration can be implemented. The ‘peaking through’ advantage does not occur so much in a close-packed field with polar/elevation
tracking.
Clearly, ‘more optically efficient’ layouts can be generated for energy
production at some defined design point; however, these may not be the
most cost effective, and often do not consider receiver constraints.
© Woodhead Publishing Limited, 2012
278
Concentrating solar power technology
Ease of access for maintenance
It is important to maintain ready access to each heliostat to allow regular
washing and occasional repairs. The RS configuration allows this, as does
the east-west, while some proposed layout schemes are exceedingly close
packed or essentially random (each heliostat is installed at the ‘best position’ based on area left by previous heliostats). Pragmatic consideration
needs to be given by field designers to installation and ongoing O&M
requirements, and some minimum constraints are needed prior to optical
optimization.
8.9
Future trends
R&D in research institutions and in industry will lead to better, longerlived reflectors and absorbers, which will reduce the size and cost of systems
delivering a specific power proportionately. Mass production and learning
will reduce the manufacturing cost and materials requirements and likely
follow a typical 85% learning curve (cost drops by 15% for every doubling
of installed plant) until irreducible materials costs dominate. Well-designed
small systems can compete with larger systems if they are clustered to
achieve economy of scale, co-located with an industrial partner, or designed
for unattended operation. Otherwise, operator costs will exceed the income.
The ability to operate efficiently with large, low-cost storage is a big
advantage of CSP. For central receivers, storage efficiency is higher than in
lower temperature systems (storage systems are covered in detail in Chapter
11). Of the numerous thermal storage options under development, two-tank
molten salt at this time appears to be the most attractive within its temperature constraints. Thermocline storage is more complicated to operate
and maintain and less efficient, while phase-change storage shows promise
though a deal of work is required for temperature ranges of interest to
central receivers. Both do have the potential cost advantage of requiring
only one tank. Solar multiples of 2–4 will be common, with most of the
energy being sent to storage to provide dispatchability and satisfy peaking
loads or in the far term, perhaps base loads. Molten salt was used successfully in Solar Two, and is in current use at Gemasolar, and improved salts
are being developed to increase the outlet temperature well above 565°C
and to simultaneously reduce the freezing temperature to well below 200°C.
This would allow higher efficiency supercritical turbines to be used, and
heat tracing needs will be reduced.
High temperature volumetric receivers using air or a gas as a working
fluid are under development, and can be used with Brayton cycle engines
at high efficiency. Combined with a ceramic bed storage unit, these show
promise. The combination of a large fixed focal zone and of high achievable
temperatures also makes the CR system particularly suitable for
© Woodhead Publishing Limited, 2012
Central tower concentrating solar power (CSP) systems
279
solar driven chemical processes such as sustainable fuels production (see
Chapter 20).
Overall, central receiver systems are on the brink of a major growth
phase. Many hundreds of 100+ MWe plants, many with capacity factors well
over 50%, are likely to be built worldwide in the next 50 years providing a
significant proportion of world electricity demand, and possibly fuels. Most
will be dry-cooled. Such a typical plant will produce over 400 GWhr of
electricity per year, with essentially zero on site pollution.
As central receiver plants demonstrate they are basically benign and less
risky than the alternatives that deface land and pollute water, emit pollutants or leave a legacy of nuclear waste, the desert and arid areas of the
world will bloom with central receiver plants and there will be plentiful,
clean, inexhaustible, and inexpensive energy for many generations to come.
8.10
Sources of further information and advice
Rabl A (1985), Active solar collectors and their applications, New York and
Oxford: Oxford University Press
An excellent general reference and detailed textbook for all solar things
thermal, from equipment, solar geometry and insolation, to flat plate and
concentrating collectors, heat transfer, modelling, economics, and
optimization.
Boer K W (ed.) (vol. 1–12), Goswami Y, Boer K W (eds) (vol. 13–14),
Goswami Y (ed.)(vol. 15–17), Advances in Solar Energy, Boulder, CO:
American Solar Energy Society
Goswami Y, ed. (vol. 18- ) Advances in Solar Energy, Freiburg, Germany:
International Solar Energy Society
This excellent series contains chapters by experts in the general field of
solar energy. Vol. 10 contains a chapter by S. Awerbuch on valuing the economics of solar collectors. Vol. 13 contains several chapters on climate
change, carbon limitation, and prospects for solar to mitigate the effects.
Vol. 15 has a chapter on the design, operation and performance of the SEGS
parabolic trough plants and a chapter providing a ‘recipe’ for the design
and optimization of central receiver plants. Vol. 16 contains a chapter on
recent advances in solar trough technology. Vol. 17 discusses solar heat for
industrial processes and issues related to solar energy in the Middle East
and North Africa.
C.-J. Winter, R. L. Sizmann and L. L. Vant-Hull (eds) (1991) Solar power
plants, Berlin and New York: Springer Verlag
The editors gathered a group of authors who were all currently active
workers in the field to collect vital information on a common basis for the
design, operation, costing, and performance of both concentrating solar
© Woodhead Publishing Limited, 2012
280
Concentrating solar power technology
thermal plants and photovoltaic plants. The book features an extensive
tabulation of solar thermal and of photovoltaic power plants and test facilities around the world (as of 1991).
Journal of Solar Energy Engineering, American Society of Mechanical
Engineers
ASME Solar Energy Proceedings
Solar Energy, International Solar Energy Society, Elsevier
ISES bi-annual conference proceedings
Each of these publications contain significant archival articles on the state
of the art in many aspects of solar, including CSP.
Vant-Hull L L (1985), ‘Solar thermal power generation’, Natural Resources
Journal, 25, 1099–1117
Vant-Hull L L(1992), ‘Solar thermal electricity: an environmentally benign
and viable alternative’, Perspectives in Energy, 2, 157–166
Burkhardt, J J III, G Heath and E Cohen (2012), ‘Life cycle greenhouse gas
emissions of trough and tower concentrating solar power electricity generation: systematic review and harmonization’, Journal of Industrial
Ecology, 16 (S1), S93–S109, www.wileyonlinelibrary.com/journal/jie
The above three papers present energy input vs energy output results presented as life cycle analysis (LCA); important as a measure of CO2 mitigation. The first two present a detailed evaluation for the commercial designs
on which Solar One and Solar Two were based. The third is a meta-analysis
in which many trough and central receiver LCAs are ‘harmonized’ to put
them on an equal footing regarding assumed insolation, etc., and a ‘best
value’ is defined for each (nearly the same).
SolarPACES Proceedings
In 1982, a group of designers and operators of concentrating solar thermal
power plants was assembled in Claremont, CA, from around the world by
the US DOE to discuss the state of the art. Conferences have been held
every two years since then (annually since 2006), and the proceedings constitute a significant history and research direction for the field. In about
1998 the title ‘Solar Power and Chemical Engineering Systems’ (SolarPACES) was adopted by the organization. SolarPACES is now an IEAsponsored organization with a series of tasks assigned to various country
task leaders to organize. Each country is responsible for funding in-house
research on projects it accepts.
L L Vant-Hull, Heliostat Field Analysis, May 1978 ORO 5178-78-2: UC 62
A contractor’s report on a sensitivity analysis of the 100 MWe Solar One
baseline plant vs heliostat cost, land cost, land slope, receiver size, tower
height, etc. Note: at 10 MWe, Solar One was a scaled prototype of this
100 MWe design.
© Woodhead Publishing Limited, 2012
Central tower concentrating solar power (CSP) systems
281
F W Lipps and L L Vant-Hull, Parametric study of optimized central
receiver systems, Proceedings of the 1978 Annual Meeting, Vol. 2.1,
American Section of the International Solar Energy Society, Inc.
pp. 793–798.
F K Falcone, A Handbook for Solar Central Receiver Design,
SAND80-8000.
L G Radosevich, Final report on the power production phase of the 10-MWe
solar thermal central receiver pilot plant (Solar One), SAND87-8022.
B Kelly, Lessons Learned, Project History and Operating Experience of the
Solar Two Project, SAND2000-2598.
A B Zavoico, Solar Power Plant Design Basis Document,
SAND2001-2100.
J E Pacheco (ed.) Final Test and Evaluation Results from the Solar Two
Project, SAND2002-0120.
G J Kolb, An Evaluation of Possible Next-Generation High-Temperature
Molten-Salt Power Towers, SAND2011-9320.
The above documents from Sandia staff and contractors represent a definitive report on the information gained during the design, construction, operation, and evaluation of the 10 MWe emulation of a 100 MWe salt cooled
power plant with molten salt storage.
8.11
Acknowledgements
The author would like to thank the editors for extra help with some sections
of this chapter.
8.12
References
Arizona Power Systems (2006), ‘APS completes first solar trough power plant in
Arizona’. Available from: http://www.aps.com/main/news/releases/release_315.
html (accessed 10 February, 2012).
Baum V A, R R Aparase, and B A Garf (1957), ‘High-power solar installations’,
Solar Energy, 1(1), 6–12.
Bradshaw R W (1987), ‘Oxidation and chromium depletion of alloy 800 and 316 SS
by molten NaNo3-KNO3 at temperatures above 600 deg C,’ Technical Report
SAND86-9009, SANDIA National Laboratories, Livermore, CA.
Chen Y T, K K Chong, T P Bligh, L C Chen, J Yunus, K S Kannan, B H Lim, C S
Lim, M A Alias, N Bidin, O Aliman, S Salehan, S A H Shk Abd Rezan, C M Tam,
and K K Tan (2001), ‘Non-imaging focusing heliostat’, Solar Energy, 71(3),
155–164.
Dellin T A and M J Fish (1979), ‘Heliost at design cost/performance trade offs’,
SAND-79-8248.
DoE (1977), ‘Recommendations for the conceptual design of the Barstow,
California, Solar Central Receiver Pilot Plant, Executive Summary’, SANDIA
© Woodhead Publishing Limited, 2012
282
Concentrating solar power technology
Albuquerque and Livermore, for the United States Energy Research and Development Agency, October, Contract AT (29-1) 789, SAND77-8035.
Eskom (2002), A study for a 100 MWe central receiver in SA carried out by the
Bechtel – affiliated company, NEXANT under funding from the South African
utility, Eskom. No external documents have been published.
eSolar (2011), ‘Our solution’ Available from: http://www.esolar.com/our_solution/
(accessed December, 2011).
Francia G (1968), ‘Pilot plants of solar steam generating systems’, Solar Energy, 12,
51–55.
Hildebrand A F and L L Vant-Hull (1977), ‘Power with heliostats’, Science, 198,
1139–1146.
Hildebrand A F, G M Haas, W R Jenkins, and J P Colaco (1972), ‘Large-scale concentration and conversion of solar energy’, EOS, 53, 684–692.
Ho C K (2008), ‘Software and codes for analysis of concentrating solar power technologies’, SAND 2008-8053.
Kattky K and L Vant-Hull (2012), ‘Optimum target reflectivity for heliostat washing’,
paper submitted to SolarPACES 2012 Symposium.
Kelly B D (2000), ‘Lessons learned, project history and operating experience of the
Solar Two project’, SAND2000-2598. Sandia National Laboratories, Albuquerque,
NM.
Kelly B D (2010), ‘Advanced Thermal Storage for Central Receivers with Supercritical Coolants’, DOE Grant DE-FG36-08GO18149, Abengoa Solar Inc., Lakewood, CO, June 15.
Kisler B L (1986), A users’ manual for DELSOL3: a computer code for calculating
the optical performance and optimal system design for solar thermal central receiver
plants, SAND 86-8018, SANDIA National Laboratories, Livermore, CA.
Kolb G J (2011), ‘An Evaluation of Possible Next-Generation High-Temperature
Molten-Salt Power Towers’, SAND2011-9320. Sandia National Laboratories,
Albuquerque, NM.
Kolb G J, S A Jones, M W Donnelly, D Gorman, R Thomas, R Davenport, and R
Lumia (2007), ‘Heliostat Cost Reduction Study’, SAND2007-3293. Sandia
National Laboratories, Albuquerque, NM, June.
Lata J, S Alcalde, D Fernández, and X Lekube (2010), ‘First Surrounding Field of
Heliostats in the World for Commercial Solar Power Plants – Gemasolar’. SolarPACES 2010 Symposium, Perpignan, France.
Lipps F W and L L Vant-Hull (1978), ‘A cellwise method for the optimization of
large central receiver systems’, Solar Energy, 20, 505–516.
Lovegrove K and A Luzzi (2002), ‘Solar thermal power systems’, Encyclopedia of
Physical Science and Technology, 3rd edn, Volume 15, Academic Press, San Diego,
CA, pp. 223–235.
Mancini, T R, J A Gary, G J Kolb, and C K Ho (2011), ‘Power tower technology
roadmap and cost reduction plans’, SAND 2011-2419.
Meduri P, C Hannsmann, and J Pacheco (2010), ‘Performance characterization and
operation of eSolars’s Sierra suntower power tower plant’, SolarPACES 2010
Symposium, Perpignan, France.
Pacheco J, M C Moursund, D Rogers, and D Wasyluk (2011), ‘Conceptual design of
a 100 MWe modular molten salt power tower plant’, SolarPACES 2011 Symposium, Granada, Spain.
© Woodhead Publishing Limited, 2012
Central tower concentrating solar power (CSP) systems
283
Pitman C L and L L Vant-Hull (1986), ‘Performance of optimized solar central
receiver systems as a function of receiver thermal loss per unit area’, Solar Energy,
37(6), 457–468.
Rabl A (1985), Active solar collectors and their applications, New York, Oxford,
Oxford University Press.
Smith D (1992), ‘Design and optimization of tube type receiver panels for molten
salt applications’, Solar Engineering – Proceedings of the 1992 ASME International Solar Energy Conference, Vol. 2, pp. 1029–1036.
Solar Tres (2000), Task 1, ‘Application of solar two lessons learned to a commercial
plant’, Cooperative Agreement DE-FC04-01AL67310, prepared by Nexant for
the United States Department of Energy Albuquerque Operations Office and
Sandia National Laboratories, Albuquerque, New Mexico.
Stein W (2011), Private communication, CSIRO, Australia.
Stoddard M C, S E Faas, C J Chiang, and J A Dirks (1987) ‘Solergy – A computer
code for calculating the annual energy from central receiver power plants’,
SAND86-8068. Sandia National Laboratories, Albuquerque, NM.
Trombe F (1957), ‘Solar furnaces and their applications’, Solar Energy, 1(2–3), 9–15.
Utility Study (1988), Arizona Public Service, ‘Utility Solar Central Receiver Study,
Vols. 1 & 2, Arizona Public Service (APS), Black & Veatch Engineers-Architects
(BV), Babcock & Wilcox (B&W), Pitt-Des Moines, Inc. (PDM), Solar Power
Engineering Co. (SPECO), and University of Houston (UH)’, November, DOE
Reports No. DOE/AL/38741-1 and 38741-2 (NTIS).
Vant-Hull L L (1991), ‘Concentrator optics’, in C Winter -J, R L Sizmann and L L
Vant-Hull, Solar Thermal Power Plants, New York, Berlin, Heidelberg: Springer
Verlag.
Vant-Hull L L and Pitman C L (1990), ‘Static and Dynamic Response of a Heliostat
field to Flux Density Limitations on a Central Receiver’, Proceedings of 1990
ASME Intern. Solar Engineering Conference, Miami FL.
Vant-Hull L L, M E Izygon, and C L Pitman, (1996), ‘Real Time Computation and
Control of Solar Flux Density on a Central Receiver (Solar Two – Preheat)’, Solar
Engineering 1996, presented at 1996 ASME International Solar Energy Conference, San Antonio, TX.
Walzel M D (1978), ‘An investigation of optimum heliostat spacings for the subtower region of a solar power plant’, Proceedings of the 1978 annual meeting, Vol.
2.1, American section of the International Solar Energy Society, Inc.
Walzel M D, F W Lipps, and L L Vant-Hull (1977), ‘A solar flux density calculation
for a solar tower concentrator using a two-dimensional Hermite function Expansion’, Solar Energy, 19(3), 239–253.
Zaibel R, E Dagan, J Karni, and H Ries (1995), ‘An astigmatic corrected targetaligned heliostat for high concentration’, Sol. Energy Mater. Sol. Cells, 37,
191–202.
© Woodhead Publishing Limited, 2012
9
Parabolic dish concentrating solar power
(CSP) systems
W. S C H I E L and T. K E C K,
schlaich bergermann und partner, Germany
Abstract: The main parts and working principle of dish engine (dish
Stirling) systems are explained. An overview of the historical
development and present systems is given. The energy conversion
processes are explicated as well as performance and operational
characteristics. Manufacturing aspects of components are discussed and
future development trends are shown.
Key words: dish concentrator, parabolic concentrator, dish/Stirling.
9.1
Introduction
Dish concentrating solar power (CSP) systems use paraboloidal mirrors
which track the sun and focus solar energy into a receiver where it is
absorbed and transferred to a heat engine/generator or else into a heat
transfer fluid that is transported to a ground-based plant. Dish concentrators have the highest optical efficiencies, the highest concentration
ratios and the highest overall conversion efficiencies of all the CSP technologies. The bulk of commercial CSP activity with dish concentrators
involves the use of receiver integrated Stirling engines for direct production
of electricity. However, dish concentrators can be used to drive the whole
range of energy conversion processes that are open to CSP technologies in
general.
The field of possible applications covers, on the one hand, the support of
smaller or large grid connected systems, and on the other hand, stand-alone
systems that can power, for example, water pumps or desalination plants. If
dish Stirling systems are installed in clusters, applications up to 10 MW can
be realized. Above this range, other solar thermal systems may be economical or more efficient.
Dish Stirling systems have demonstrated the highest efficiency of any
solar power generation system by converting nearly 30% of direct normal
incident (DNI) solar radiation into electricity after accounting for parasitic
power losses (EPRI Report, 1986). These high-performance solar power
systems have been in development for more than two decades, with the
primary focus in recent years on reducing the capital and operating costs
284
© Woodhead Publishing Limited, 2012
Parabolic dish concentrating solar power (CSP) systems
285
of systems. Dish Stirling systems currently cost about US$10,000 per kW
installed; major cost reduction will occur with mass production and further
development of the systems. Substantial progress has been made to improve
reliability, thereby reducing the operating and maintenance (O&M) costs
of the systems.
As capital costs drop to about US$3,000 per kW, promising market
opportunities appear to be developing in green power and distributed generation markets in the south-western United States, India, the Mediterranean region as well as in southern Europe and Africa.
With the worldwide restructuring of utility markets, the emergence of
green power markets, and the increased worldwide demand for distributed
generation, the opportunities for small power systems ranging in size from
a few kW to several MW are increasing at a rapid rate. This increasing
demand is largely being met today by existing internal combustion and gas
turbine power generators, but it is also the motivation for new technology
development such as micro turbines, fuel cells, and other alternative power
generators. Large-scale grid connected systems based on dish systems with
either Stirling engines, Brayton cycles or steam generation for groundbased turbines are also proposed. In this regard, dish systems are the least
commercially developed CSP technology today. However, the high conversion efficiencies achievable motivate the continued efforts.
9.2
Basic principles and historical development
9.2.1
Basic principles
A dish system consists of: (a) a parabolic shaped concentrator, (b) tracking
system, (c) solar heat exchanger (receiver), (d) an (optional) engine with
generator and (e) a system control unit (Fig. 9.1). The concentrator tracks
the sun bi-axially in such a way that the optical axis of the concentrator
always points to the sun. The solar radiation is focused by the parabolic
concentrator onto the solar receiver which is situated close to the focal
point of the parabola. The receiver captures the high temperature thermal
energy into a fluid that is either the working fluid for a receiver-mounted
engine cycle, or is used to transport the energy to ground-based processes.
In the case of a receiver-mounted engine, a directly coupled generator
finally converts mechanical energy into electricity.
A reflective surface on the paraboloidal concentrator, either metallized
glass or plastic, reflects incident sunlight to a small region called the focus.
The ideal shape of the reflecting surface of a solar concentrator is a paraboloid. The size of the focus depends on the precision of the shape of the
concentrator, surface reflectivity and condition as well as focal distance.
Common dish concentrators achieve geometric concentration ratios
© Woodhead Publishing Limited, 2012
286
Concentrating solar power technology
1
3
11
4
2
10
9
5
7
8
Key
1 Concentrator bowl
2 PCU supporting structure
3 Power conversion unit (PCU)
4 Elevation drive guide
5 Azimuth drive guide
6 Foundation
7 Azimuth drive
8 Electrical cabinet
9 Pedestal
10 Annular mounting structure
11 Elevation bearing
6
9.1 Schematic representation of an example of a dish system.
between 1,500 and 4,000. Dish concentrator diameters range from 1–2 m
up to 25 m.
In order to track the sun, concentrators must be capable of moving about
two axes. Generally, there are two ways of implementing this, both having
advantages:
•
•
Azimuth-elevation tracking (illustrated in Fig. 9.2a), in which the dish
rotates in a plane parallel to the earth surface (azimuth) and around an
axis perpendicular to it (elevation), gives the collector up/down and left/
right rotations. Rotational rates about both axes vary throughout the
day but are predictable.
Alternatively with the polar-equatorial tracking method (illustrated in
Fig. 9.2b), the collector rotates about an axis parallel to the Earth’s axis
of rotation. The collector rotates at a constant rate of 15°/hr, the same
rotation rate as the Earth’s. The other axis of rotation, the declination
axis, is perpendicular to the polar axis. Movement about this axis occurs
slowly and varies by ±23½° over a year (a maximum rate of 0.016°/hr).
The biaxial tracking system is normally driven by electric motors working
through gearbox units, although hydraulic systems have also been developed. The tracking position is found with sun or reflected beam sensors and/
or with a tracking algorithm that calculates the actual position of the sun
from the date and time of day for the known location of the system. For
© Woodhead Publishing Limited, 2012
Parabolic dish concentrating solar power (CSP) systems
287
(a)
(b)
9.2 Principle of (a) azimuth-elevation and (b) polar-equatorial mounted
systems.
© Woodhead Publishing Limited, 2012
288
Concentrating solar power technology
control of these drives and for the whole system including the Stirling
engine, micro controllers or PCs are used. The operation is therefore fully
automatic and remote control via the Internet is possible.
The receiver has two functions: (1) to absorb as much of the solar radiation reflected by the concentrator as possible and (2) to transfer this energy
as heat to the working fluid.
Although a perfectly reflecting paraboloid reflects parallel rays to a point,
the sun’s rays are not quite parallel because the sun is not a point source.
Also, any real concentrator is not perfectly shaped. Therefore, concentrated
radiation at the focus is distributed over a small region, with the highest
concentration of flux in the centre, decreasing towards the edge.
Efficient receivers for dish systems are cavity receivers with a small
opening (aperture) through which concentrated sunlight enters. The
absorber is placed behind the aperture to reduce the intensity of concentrated solar flux. The insulated cavity between the aperture and absorber
reduces the amount of heat lost. The receiver aperture is optimized to be
just large enough to admit most of the concentrated sunlight but small
enough to limit radiation and convection loss (Stine and Harrigan, 1985).
9.2.2
Historical development
As with the other approaches to concentrator design, the concept of using
mirrored dishes to focus the sun has been around since 200 BC. One of the
earliest actual implementations of a dish system was by the Frenchman
Augustin Mouchot, who built a series of dish-driven engine systems as early
as 1864, and displayed a dish concentrator at the Universal Exhibition in
1878 in Paris (Gordon, 2001).
The dish technology of today began evolution with the sudden increase
in activity in all aspects of CSP in the early 1980s following the oil price
shock in the 1970s. Much of the early activity was centred in the USA.
Innumerable dish prototypes were designed and built by researchers, large
and small commercial organizations and even private individuals. Stine and
Diver (1994) have edited a comprehensive overview of systems as well as
components. A representative selection of key examples is reviewed here.
The first of the commercially prototyped systems using a Stirling engine,
the 25 kW Vanguard system built by ADVANCO in Southern California,
achieved a reported world record net solar-to-electric conversion efficiency
of 29.4% (EPRI Report, 1986). The Vanguard dish Stirling system utilized
a glass-faceted dish 10.5 m in diameter, a direct insolation receiver (DIR),
and a United Stirling 4-95 Mark II double-acting kinematic Stirling engine
(Fig. 9.3).
In 1984, two 50 kW dish Stirling systems were built, installed and operated in Riyadh, Saudi Arabia, by schlaich bergermann und partner (sbp) of
© Woodhead Publishing Limited, 2012
Parabolic dish concentrating solar power (CSP) systems
289
9.3 Vanguard 1 concentrator.
9.4 17 m metal membrane concentrator built by sbp.
Stuttgart, Germany (Koshaim, 1986). A similar unit was also installed at the
German Aerospace Centre (DLR) facility in Lampoldhausen, Germany.
The dishes were 17 m diameter stretched-membrane concentrators, by
drawing a vacuum in the plenum space formed by the dish rim and front
and back thin steel membranes. The optical surface of the dish was made
by bonding thin glass tiles to the front membrane. The receivers for the sbp
dishes were direct illuminated tube receivers and the engines were United
Stirling 4-275 kinematic Stirling engines (Fig. 9.4).
A dish Stirling system was built by McDonnell Douglas Aerospace Corporation (MDAC) in the mid-1980s and, when MDAC discontinued development of the technology, the rights to the system were acquired by
Southern California Edison (SCE) (Lopez and Stone, 1992, 1993). The parts
© Woodhead Publishing Limited, 2012
290
Concentrating solar power technology
9.5 McDonell Douglas Corporation concentrator (copyright: DOE/
NREL).
for eight systems were manufactured, and three systems were tested in the
early 1980s. The MDAC/SCE dish was the first dish Stirling system designed
to be a commercial product (Fig. 9.5). It was built on the design of the
Vanguard dish Stirling system, using the same DIR and the USAB 4-95
Mark II engine. SCE operated the system from 1985 to 1988. Stirling Energy
Systems (SES) of Phoenix, Arizona, acquired the technology rights and
system hardware in 1996 and have continued development of the system.
In 1989, schlaich bergermann und partner built the first of their smaller
7.5 and 8.5 m stretched-membrane concentrators equipped with a 10 kW
SOLO V160 Stirling engine. First, in polar tracking configuration and later
in an azimuth-elevation tracking configuration, six of the systems have
together operated for more than 30,000 hr in sun at the Plataforma Solar
de Almería in Southern Spain (Fig. 9.6).
In 1991, Cummins Power Generation, working under cost share agreements with the US Department of Energy and Sandia National Laboratories, started the development of two dish Stirling systems – a 7 kW system
for remote applications and a 25 kW system for grid-connected power
generation (Fig. 9.7) (Gallup and Mancini, 1994; Bean and Diver, 1995).
Cummins was innovative in its dish Stirling systems, incorporating advanced
technologies into the designs, such as a solar concentrator with a polar-axis
© Woodhead Publishing Limited, 2012
Parabolic dish concentrating solar power (CSP) systems
291
9.6 7.5 m Generation I, polar tracking metal membrane dish systems
built by sbp operating at PSA, Spain (Copyright: schlaich bergermann
und partner).
9.7 Cummins Power Generation CPG-460 concentrator.
drive and polymer, stretched-membrane facets, heat-pipe receivers, and
free-piston Stirling engines. The heat-pipe receiver transfers the absorbed
solar heat to the engine by evaporating sodium and condensing it on the
tubes of the engine heater head. The receiver serves as a thermal buffer
between the concentrator and the engine, and because it transfers heat to
the engine by condensation, it allows the engine to operate at a high and
uniform average temperature and efficiency (Andraka et al., 1993). The two
Cummins programmes made progress, but were terminated in 1996 when
© Woodhead Publishing Limited, 2012
292
Concentrating solar power technology
Cummins’ parent company, Cummins Engine Company, realigned business
along its core area of diesel engine development. The assets of the Cummins
solar operations were sold to Kombassan, a holding company in Alanya,
Turkey (Mancini et al., 2003).
In these early projects, dish Stirling systems have demonstrated their
capability of producing electricity for the grid and for remote power applications at high solar to electric efficiencies. All systems so far were built in
single piece production and therefore have a high investment cost level.
Additionally, Stirling motors still require regular maintenance. Thus the
aims for further developments are an increase in system reliability and
further cost reduction. Therefore current efforts are focused on establishing
reliability and, through break-and-repair approaches, identifying the components that require improvement, redesign and replacement. In a parallel
approach, advanced components, such as system controls and improved
optical surfaces, that promise higher efficiencies and reliabilities at lower
cost, are being developed and tested. In addition, industrial series component production is being implemented.
Along with the approach of using Stirling engines with receivers mounted
in the dish focal point, the use of dishes to provide heat to a central generating plant has also been pursued. Many receiver development experiments
of various kinds have been reported in the research literature.
A notable example of an attempt to demonstrate commercial scale operation was the La Jet ‘Solarplant1’ 4.9 MWel system built in California in
1984. As shown in Fig. 9.8, it consisted of 700 La Jet dishes (essentially the
same as the Cummins technology) with stretched-membrane mirror elements. The total collecting area was 30,590 m2. Six hundred of the dishes
produced saturated steam at approximately 6 MPa and the remaining 100
dishes were used to further superheat it to 460°C. The plant was operated
9.8 La Jet plant with 700 units in California.
© Woodhead Publishing Limited, 2012
Parabolic dish concentrating solar power (CSP) systems
293
9.9 ANU dish installation in White Cliffs, Australia (Copyright: ANU
Solar Thermal Group).
until 1990. Whilst this plant successfully demonstrated centralized steambased generation with dishes, the main negative issues were a lack of durability with the mirror membrane reflective polymer film and a large startup
time due to excessive thermal inertia in the receivers.
In Australia, the early work of the group at the Australian National University (ANU) leads to the construction of a 14 dish system in the remote
town of White Cliffs in New South Wales (Australia) (Fig 9.9). Each dish
has an aperture area of 20 m2 and has small flat mirror tiles bonded to a
single fibreglass paraboloid. Superheated steam was generated directly in
monotube ‘semi cavity’ receivers and networked to a central power block,
using a 25 kWel reciprocating steam engine/generator (Kaneff, 1991).
In 1999 and 2000, WG Associates designed the WGA-500 dish concentrators MOD1 and MOD2 with 8.8 m diameter and 41 m2. The dish was pylon
mounted, the reflector made from sandwich panels with thin glass mirrors.
Two prototypes were built and operated with the 10 kW SOLO 161 Stirling
engine.
9.3
Current initiatives
In this section a short description of the ongoing major dish initiatives is
given.
9.3.1
Stirling Energy Systems (SES)
In 1996, Stirling Energy Systems, Inc. (SES) was formed and acquired all
design and engineering patents on the solar Stirling dish engine technology
© Woodhead Publishing Limited, 2012
294
Concentrating solar power technology
9.10 Stirling Energy Systems new 25 kW concentrator (SunCatcher)
(Copyright: Sandia National Laboratories/Randy Montoya).
that had been developed over nearly three decades by MDAC/SCE. SES
launched two strategic, collaborative public–private partnerships with
Sandia National Laboratory, Albuquerque, New Mexico and the United
States Department of Energy (DOE) in order to commercialize this technology. Over this period, SES has re-engineered the licensed technology,
improved its design, performance and cost, while achieving high solar-toelectricity conversion efficiency. Today, SES technology is a dish Stirling unit
called SunCatcher using a double-acting 25 kW kinematic Stirling engine
based on the developments by United Stirling in Sweden (Fig. 9.10). A plant
with 60 dishes, 1.5 MWel system using the SunCatchers has been constructed
by SES in Pheonix, USA. SES together with the sister company, Tessera
Solar North America, has signed several power purchase agreements (PPA)
for two large plants with 709 and 850 MW with utilities in the US. Both the
system and the project development for the commercial plants had been
far advanced, but SES went into bankruptcy after failing to receive a governmental loan guarantee.
9.3.2
schlaich bergermann und partner (sbp)
In 1998 sbp started together with European partners the development of
the EuroDish (Fig. 9.11). In a first step, two 8.5 m diameter dish concentrators, equipped with an improved Stirling engine, were erected and tested at
the Plataforma Solar de Almería (PSA). The EuroDish incorporates a
newly developed concentrator, made up of a sandwich shell from fibre glass
reinforced plastic and the well proven and further improved single-acting
© Woodhead Publishing Limited, 2012
Parabolic dish concentrating solar power (CSP) systems
295
9.11 10 kW EuroDish by sbp.
SOLO Stirling 161 with 10 kWel capacity. The tracking and control system
was also revised and simplified, and remote control capability has been
implemented. Two EuroDish prototypes were built in Spain for testing and
continuous operation.
After intensive testing several so-called ‘country reference units’ were
erected in Spain, Germany, France, Italy and India to demonstrate the
technology to the market and gain substantial operation experience at
various sites in the world under different climatic conditions.
9.3.3
Infinia Corporation
Infinia Corporation, based in Ogden, Utah, USA is a privately owned technology company that has been developing free piston Stirling engines since
1967. In 2006 Infinia started the development of the 4.7 m diameter
PowerDish equipped with a self-developed, low-cost, long-life and
maintenance-free 3.2 kW free piston Stirling engine. The dish was designed
together with schlaich bergermann und partner. It is an automatic, selfcontained system and Infinia claims no maintenance of the hermetically
sealed engine is required over the whole life span of 25 years.
A first prototype was erected in 2007. Pre-production and test units are
currently operating at ten different sites around the world; Infinia commissioned their first commercial installation of 30 units in Yuma, Arizona in
August 2010 (Fig. 9.12). Infinia has partnered with large Tier 1 automotive
component manufacturers and suppliers and has utility-scale projects in the
US, Europe and India following soon after the project in Yuma with full
production launch in 2012.
© Woodhead Publishing Limited, 2012
296
Concentrating solar power technology
9.12 PowerDish installation in Yuma, Arizona 2010 by Infinia Corp.
(Copyright: Infinia Corp.).
9.13 HelioFocus 500 m2 dish.
9.3.4
HelioFocus
HelioFocus Ltd of Ness Ziona, Israel, was founded in 2007. They completed
a low cost, large scale dish development (500 m2) together with schlaich
bergermann und partner (Fig. 9.13). The first prototype was erected in mid
2011 as part of a solar boosting experiment with the Israeli utility company.
The dish is made of a flat support structure with mirrors arranged in a
Fresnel-like array and tracks the sun using a hydraulic drive system. The
first application is generating high temperature air as heat transfer fluid
© Woodhead Publishing Limited, 2012
Parabolic dish concentrating solar power (CSP) systems
297
(HTF), using a pressurized volumetric receiver. Steam is produced via a
heat exchanger and fed into a fossil-fuelled power plant (boosting). In the
medium term, HelioFocus intend to develop a system with a micro turbine.
9.3.5
Solar Cat/SouthWest Solar
SouthWest Solar Technologies of Phoenix, Arizona, USA, have also developed a large dish concentrator, measuring 23 m in diameter (320 m2). The
prototype was commissioned in 2011. It is suspended on a pylon and feeds
a 80 kWel micro turbine from Brayton Energy LLC. Hybrid operation and
compressed air storage is intended (SouthWest Solar, 2011).
9.3.6
Solar Systems
The Australian-based company Solar Systems Pty. Ltd, now owned by Silex
Systems Ltd, has been working in CPV with dish concentrators since the
late 1990s. Their CS500 130 m2 dish generates 35 kW and is pylon mounted
(Fig. 9.14). Several projects with a total of 40 units have been realized.
Today, the system is called ‘Dense Array Converter’, with a similar dish
design measuring 140 m2 and a PV generator with 40% efficiency. According to Silex information, a 60 unit/2 MW plant shall be commissioned in
early 2013 in Mildura, and another 102 MW (40 kW per dish) will follow.
9.3.7
Australian National University (ANU)
Following on from the 14-dish system at White Cliffs, in 1994, the 400 m2
dish SG3 was designed and built by the Australian National University
9.14 Solar Systems CS500 Dense Array CPV dish system.
© Woodhead Publishing Limited, 2012
298
Concentrating solar power technology
9.15 ANU 400 m2 dish SG3 (Copyright: ANU Solar Thermal Group).
(ANU) (Fig. 9.15). It is made up of 54 triangular mirror facets on a space
frame. The concentrator is mounted on a turntable and rotates in azimuth
on a base frame with six wheel assemblies on a concrete ring. A monotube
boiler receiver was used to generate superheated steam. The size was motivated by an analysis that concluded that larger dishes are more cost effective per unit area than small ones. Large-scale grid connected systems using
ground steam turbine-based generation were targeted, so the availability of
suitable engines to be mounted at focal did not constrain the design.
In 2009, a dish design with 500 m2 aperture area was designed and built
by ANU in collaboration with Wizard Power Pty Ltd, a startup company
established to commercialize the technology. The new dish was optimized
for mass production for large-scale plants. It features 380 interchangeable
square mirror panels which are also designed to provide a structural contribution for the dish (Fig. 9.16). The panels are again supported by a space
frame and mounted on a turntable running on wheels on a steel track
(Lovegrove et al., 2011).
9.3.8
Others
A considerable number of small and early stage initiatives both in research
centres and the commercial arena are developing dish and dish/engine
systems, too numerous to be mentioned here. They cover a large variety of
designs and size.
9.4
Energy conversion, power cycles and equipment
Although a Brayton engine has been tested on a dish (Jaffe, 1988) and some
companies are currently adapting micro-turbine technology to dish engine
© Woodhead Publishing Limited, 2012
Parabolic dish concentrating solar power (CSP) systems
299
9.16 ANU/Wizard Power 500 m2 dish (Copyright: ANU Solar Thermal
Group).
systems, kinematic and free piston Stirling engines are currently being used
in the majority of dish systems offered commercially today. Stirling engines
are preferred for these systems because of their high efficiencies (thermalto-mechanical efficiencies in excess of 40% have been reported), high
power density (40–70 kW/litre for solar engines), and potential for longterm, low-maintenance operation. Dish Stirling systems are modular, i.e.,
each system is a self-contained power generator, allowing their assembly
into plants ranging in size from a few kilowatts to tens of megawatts.
9.4.1
Stirling engines
Thermal energy provided by concentrated solar radiation can be converted
into electrical energy using a Stirling engine with coupled generator. Stirling
engines belong to the group of hot-gas machines and use a closed thermodynamic process; i.e. always the same working gas is used within the working
cycle. In contrast to Otto or Diesel engines, energy is provided by external
heat supply, so that Stirling motors are also suitable for solar operation.
The basic principle of a Stirling engine is based on the cyclic compression
and expansion of gas at different temperature levels to produce a net conversion of heat energy to mechanical work. The ideal process is based on a
combination of isothermal compression of the cold and isothermal expansion of the hot medium plus constant volume (isochoric) heating and
cooling processes (Fig. 9.17(c)). Periodic temperature change – and thus
continuous operation – can be ensured by moving the working gas between
two chambers of constantly high and constantly low temperature.
For the technical realization using crankshaft-linked pistons (a kinematic
engine), a compression piston is moved to the closed side, so that the cold
© Woodhead Publishing Limited, 2012
© Woodhead Publishing Limited, 2012
(b)
Isothermal compression phase
(d)
(a)
(c)
Isothermal expansion phase
Isochorous cooling phase
Compression
cylinder
Cooler
9.17 Working principle of a Stirling engine.
Working
cylinder
Heater
Regenerator
Isochorous heating phase
Parabolic dish concentrating solar power (CSP) systems
301
working gas flows to the warm space, passing through a regenerator. The
regenerator transmits the previously absorbed heat to the working gas
(isochoric heating phase; Fig. 9.17(a)). The gas is warmed up to the temperature of the hot space while the regenerator cools down to the temperature of the cold space. Subsequently, the working gas inside the hot space
expands isothermally and absorbs the heat from the hot space (isothermal
expansion phase; Fig. 9.17(b)).
The expanding working gas moves the working piston out of the cylinder
and so performs work. When the working piston passes below dead centre
and begins to close, the hot working gas is forced to pass the regenerator
and to move into the cold space. Heat is transferred isochorically from the
working gas to the regenerator (isochoric cooling phase; Fig. 9.17(c)). The
gas is cooled down to the temperature of the cold space while the regenerator is warmed up to the temperature of the hot space. The working gas is
subsequently compressed isothermally and transmits exhaust heat to the
cold space (isothermal compression phase Fig. 9.17(d)).
The basic system components thus include the heated working cylinder,
the cooled compression cylinder and a regenerator for intermediate energy
storage. In most cases, the regenerator is a highly porous body of a high
heat capacity; this porous body has a considerably larger thermal mass
than the gas mass flowing through the body. The more complete the alternating heat transmission is performed inside the regenerator, the higher
the mean temperature difference between working and compression cylinder and thus the efficiency of the Stirling engine. If the displacing piston
is coupled to the working piston at the appropriate phase angle via a
driving mechanism or a vibratory system, the whole system can serve as
thermal engine.
In terms of mechanical design, single- and double-acting machines are
sometimes employed. In single-acting machines, only one side of the compression or expansion piston undergoes pressure fluctuations inside the
working space, while the pressure of the working gas is effective on both
sides of the piston of double-acting machines; in the latter case, they simultaneously work as compression and expansion piston.
Stirling engines can be categorized into kinematic and free piston Stirling
engines. Kinematic Stirling engines perform power transmission via a
crankshaft mechanism. A generator can be coupled to this shaft. Free piston
Stirling engines lack mechanical inter-linkage between the working piston,
the displacement device and the environment. Both pistons move freely.
The converted energy can be transferred to the exterior by an axial generator, for instance. Mechanical inter-linkage is replaced by an interior spring
damping system; this is why only two movable parts are required. The
machine can be hermetically sealed, so that tightening issues are avoided.
Free piston Stirling machines present the theoretical benefits of a simple
© Woodhead Publishing Limited, 2012
302
Concentrating solar power technology
structure and high reliability, but still lag somewhat behind in terms of
development when compared to kinematic machines.
The engines applied for dish Stirling systems use helium or hydrogen at
working gas temperatures between 600 and 800°C. Power output of the
Stirling motor is controlled by varying the working gas mean pressure or
piston stroke.
9.4.2
Brayton cycle
The Brayton engine is usually seen in the jet engine, combustion turbine or
gas turbine, as an internal combustion engine which produces power by the
controlled burning of fuel. In the Brayton engine, like in Otto and Diesel
cycle engines, air is compressed, fuel is added, and the mixture is burned.
The engine consists of a compressor turbine followed by a constant pressure
heat addition (usually combustion) and by an expansion turbine coupled
to an alternator. In a dish/Brayton system, solar heat is used to replace (or
supplement) the heat input from fuel. The resulting hot gas expands rapidly
and is used to produce power. As in the Stirling engine, recuperation of
waste heat is key to achieving high efficiency. Therefore, waste heat
exhausted from the turbine is used to preheat air from the compressor. The
recuperated gas turbine engines that are candidates for solarization have
pressure ratios of approximately 2.5, and turbine inlet temperatures of
about 850°C (1,562°F). Predicted thermal-to-electric efficiencies of Brayton
engines for dish/Brayton applications are over 30% (Koshaim, 1986; Lopez
and Stone, 1993).
9.4.3
Other cycles
As discussed previously, dish systems can also be used to provide steam for
ground-based steam turbine systems. Such systems are potentially the same
as those employed for tower, trough or Fresnel-based power generation as
discussed in Chapters 8, 7 and 6 respectively. Dishes can also be used for
PV concentrator systems as discussed in Chapter 10 and also for driving
thermochemical processes as discussed in Chapter 20.
9.4.4
Equipment
Alternator
The mechanical-to-electrical conversion device used in dish/engine systems
depends on the engine and application. Induction generators are used on
kinematic Stirling engines connected to an electric utility grid. Induction
generators synchronize with the grid and can provide single- or three-phase
© Woodhead Publishing Limited, 2012
Parabolic dish concentrating solar power (CSP) systems
303
power of either 230 or 400 volts. Induction generators are off-the-shelf items
and convert mechanical power to electricity with an efficiency of about 94%
in the relevant power capacity range.
Alternators, in which the output is conditioned by rectification (conversion to DC) and then inverted to produce AC power, are sometimes
employed to handle mismatches in speed between the engine output and
the constant electrical grid frequency. The high-speed output of a gas
turbine, for example, is converted to very high frequency AC in a high-speed
alternator, converted to DC by a rectifier, and then converted to 50 or 60 Hz
single, or three-phase power by an inverter. This approach can also have
performance advantages for operation of the engine.
Cooling system
Heat engines need to transfer waste heat to the environment. Stirling
engines use a radiator to exchange waste heat from the engine to the atmosphere. In open-cycle Brayton engines, most of the waste heat is rejected in
the exhaust. Parasitic power required for operation of a Stirling cooling
system for fan and pump, concentrator drives and controls is typically about
0.5–1 kWel for a 10 kWel system.
Receivers
In a receiver for a Stirling engine, two methods are used to transfer absorbed
solar radiation to the working gas (Fig. 9.18). In the first type of receiver,
the directly illuminated tube receiver (DIR), small tubes through which the
engine’s working gas flows, are placed directly in the concentrated solar flux
region of the concentrator. The tubes form the absorber surface (Fig. 9.19).
The other type of receiver uses a liquid metal intermediate heat transfer
fluid. The liquid metal is vaporized on the absorber surface and condenses
on tubes carrying the engine’s working gas. This second type of receiver is
called a heat pipe receiver because the vapour condenses and flows back
to be heated again (Fig. 9.20).
For receiver designs in which liquid metal is used as an intermediate heat
transfer fluid, two methods of supplying liquid metal to the absorber are
under development: pool boilers and heat pipes. With the first method, a
pool of liquid metal is always in contact with the absorbing surface. The
second method involves a wick attached to the back of the absorber.
The capillary forces in the wick draw liquid metal over the surface of the
absorber where it vaporizes.
For steam generation, monotube receivers have been developed. They
consist of a long tube that forms the absorber surface (see Fig. 9.21). The
water evaporates while flowing along the tube.
© Woodhead Publishing Limited, 2012
304
Concentrating solar power technology
Absorber
tubes
Capillary structure
Concentrated
direct radiation
Concentrated
direct radiation
Sodium vapour
Absorber
Reheater
Connecting
flange
Liquid sodium
Working
gas
inlet
Working
gas
outlet
Working gas inlet and outlet
(a)
(b)
9.18 Receiver design principle: (a) direct illuminated tube receiver; (b)
heat pipe receiver.
(a)
(b)
9.19 Example of a DIR for the 10 kW SOLO Stirling engine, section (a)
and overview (b) (SOLO).
© Woodhead Publishing Limited, 2012
Parabolic dish concentrating solar power (CSP) systems
Flameholder
Mixing tubes
305
Gas inlet
Combustion gas
recirculation
Concentrated
solar
insolation
210
Heat pipe with
wick structure
Absorber
surface
Stirling
heater
head
Stirling heat
exchanger tubes
Air
inlet
Exhaust
Air preheater
240
Heat exchanger gap
with fins
(a)
(b)
9.20 Schematic (a) and prototype (b) of a Stirling hybrid heat pipe
receiver for 10 kW engine (DLR) (Copyright: DLR).
Volumetric receivers have been designed for generation of hot air for
Brayton cycles or for thermochemical reactions (Fig. 9.22). The concentrated sunlight is absorbed in the volume of a porous high temperature
material. The heat is transferred to the air flowing through the absorber.
Usually the air is pressurized up to 10 bar which requires an airtight and
pressure-proof construction as well as a transparent window (quartz) (Buck
et al., 1996).
© Woodhead Publishing Limited, 2012
306
Concentrating solar power technology
9.21 Example of a monotube open receiver (ANU) (Copyright: ANU
Solar Thermal).
Water-cooled
target
Insulation
Inlet
Outlet
Concentrated
solar energy
Housing
Quartz window
(a)
Volumetric
absorber
(b)
9.22 ‘Example of a volumetric pressurized receiver: (a) overview and
(b) section (Copyright: DLR).
9.5
System performance
In general, overall system performance depends on system design parameters, i.e. engine efficiency and especially part load behaviour of the engine
as well as the optical performance of the concentrator (reflectivity of the
mirrors, contour accuracy of the reflector, stiffness of the support structure,
etc.) and available solar insolation (DNI). Ambient temperature acts as the
lower temperature of the thermodynamic cycle, which impacts the efficiency. This section discusses the operation and output characteristics of
typical dish Stirling systems. Other thermal power generation systems
behave in qualitatively similar ways.
© Woodhead Publishing Limited, 2012
Parabolic dish concentrating solar power (CSP) systems
307
Typical design and operation criteria are:
Operation wind speeds:
Survival wind speed:
Ambient temperature:
Humidity:
Maximum power output:
Engine pressure:
Working gas upper temperature:
Grid connection:
up to 15–20 m/s
up to 45 m/s in a wind stow position
−10 to 50°C
up to 100%
at 800–1,000 W/m2 DNI
0–15 MPa
500–700°C
single systems are connected to
230/400 V line
Due to the low thermal inertia, a dish Stirling System reacts very quickly
on changes in solar thermal input. Thus steady state operation is achieved
within a few minutes after system start.
A typical daily pattern of net electric energy production over a day is
given in Fig. 9.23 and a measured input–output diagram of a dish Stirling
system is shown in Fig. 9.24. From these diagrams it can be seen that a dish
Stirling system already starts net electric energy production when direct
beam insolation (DNI) reaches values around 200–300 W/m2 (DNI) in the
morning, depending on mechanical and thermal losses of the engine as well
as the optical performance of the concentrator. Maximum power output is
normally reached at 1,000 W/m2 of DNI. If the concentrator is over sized,
maximum power output is already achieved at a lower insolation level, e.g.,
at 800 W/m2. The negative power dip on startup of the engine is due to
warming up of the engine’s hot parts. Daily power output of a gridconnected dish Stirling system with unfavourable irradiation conditions is
shown in Fig. 9.25.
DNI (W/m2)
Power (%)
100
80
800
60
600
40
400
20
200
Power (%)
DNI (W/m2)
1000
0
–20
19:00
18:00
17:00
16:00
15:00
14:00
13:00
11:00
12:00
10:00
09:00
08:00
07:00
0
Time of the day (h)
9.23 Daily power output of a grid-connected dish Stirling system with
favourable irradiation.
© Woodhead Publishing Limited, 2012
308
Concentrating solar power technology
12
Gross capacity (kW)
10
8
6
4
In the afternoon
Start-up in the forenoon
2
0
0
100
200
300 400 500 600 700
Direct radiation (W/m2)
800
900 1000
9.24 Typical input–output diagram of a dish Stirling system.
1000
100
800
80
600
60
400
40
200
20
Power (%)
DNI (W/m2)
DNI (W/m2)
Power (%)
0
07:00
08:00
09:00
10:00
11:00
12:00
13:00
14:00
15:00
16:00
17:00
18:00
19:00
20:00
21:00
0
Time of the day (h)
9.25 Daily power output of a grid-connected Dish Stirling system with
unfavourable irradiation.
With the help of the available weather data (DNI, wind data, ambient
temperature etc.), the plant’s daily and annual net energy can be simulated
for specific locations. Using typical meteorological year (TMY2s) data,
histograms of yearly electric energy production can be developed. Modern
dish Stirling simulation codes take care of all system loss mechanisms, i.e.
engine, generator and receiver losses as well as the optical losses of the
concentrator.
Figures 9.26(a)–(c) present the results of a system simulation for a dish
Stirling power plant with 200 units each with 10 kW. All single system coefficients were considered, including dirt on the mirrors as well as clouding
© Woodhead Publishing Limited, 2012
Parabolic dish concentrating solar power (CSP) systems
309
150
100
50
0
50
12
5
20
0
27
5
35
0
42
5
50
0
57
5
65
0
72
5
80
0
87
5
95
0
10
25
Frequency (h)
Solar insolation classification
200
DNI (W/m2)
Classification of electrical energy
20
0
27
5
35
0
42
5
50
0
57
5
65
0
72
5
80
0
87
5
95
0
10
25
11
00
250
200
150
100
50
0
50
12
5
Electrical energy (MWh)
(a)
DNI (W/m2)
Electrical energy (MWh)
(b)
Monthly electric energy
450
400
350
300
250
200
150
100
50
0
1
2
3
4
5
6
7
8
9
10
11
12
(c)
9.26 Simulation of 2 MW dish/Stirling plant with 200 units of 10 kW,
showing (a) annual distribution of insolation levels, (b) electrical
energy vs. insolation level, and (c) monthly electrical energy.
(3% in the morning and evening). Availability was considered at 98%. Table
9.1 gives parameters for the expected annual energy production of the dish/
Stirling plant.
9.5.1
Hybrid operation
If power output is required independent of the existing meteorological
conditions, in the evening or at night as is required in many applications,
the dish Stirling system can, besides the use of batteries, be configured as
hybrid system. ‘Hybrid system’ means that additional fossil energy sources
© Woodhead Publishing Limited, 2012
310
Concentrating solar power technology
Table 9.1 Parameters for the expected annual energy production of a
dish/Stirling plant
Available radiation energy of direct
radiation:
Available radiation energy of direct
radiation: (aperture of a single system)
Available radiation energy of direct
radiation: (aperture of power plant)
Net annual energy production of the plant:
Net annual efficiency of the plant:
approx. 2,200 kWh/m2/a
approx. 124.9 MWh/a
approx. 25.0 GWh/a
approx. 4.5 GWh/a
approx. 18%
70
700
60
600
50
500
40
400
300
30
Burner power
20
200
10
100
Insolation (W/m2)
Power (kW)
Solar insolation
Power output
0
0
14:45 15:00 15:15 15:30 15:45 16:00 16:15 16:30 16:45 17:00 17:15
Time
9.27 Hybrid operation.
(e.g. biogas) can be used to add thermal energy to the Stirling engine to
stabilize power output over the day and during cloud passages and for
prolongation of system operation into night hours. The first hybrid systems
have already been developed and the first prototypes successfully tested
(Laing and Reusch, 1998; Moreno et al., 1999).
A system operating in hybrid mode is shown in Fig. 9.27. In hybrid mode,
the required engine pressure and consequently output power have to be
pre-set. The burner is then turned on automatically once the engine pressure falls below this level. It can be seen from the figure that the combustion
system can follow the passage of cloud very well and net electric power
output is kept fairly constant at approx. 5 kW.
© Woodhead Publishing Limited, 2012
Parabolic dish concentrating solar power (CSP) systems
9.5.2
311
Stand-alone operation
One of the possible markets for a dish Stirling system is to drive a water
pump off grid in a remote location. Standard water pump equipment is
typically driven by 480/440 V, three-phase induction motors at a constant
50 or 60 Hz. The amount of water pumped is fairly constant and is dependent on the pump and motor characteristics and the depth of the well. When
the source of the power is solar energy, the amount of water pumped should
ideally vary with the amount of sunlight available. Sandia National Labs
developed a stand-alone water pumping dish Stirling system in 2000 (Diver
et al., 2003). The following explains the approach adopted in this project.
Several options were considered to pump water independent of the utility
grid. These included coupling the induction generator with another power
source capable of providing three-phase 480 V power. The other generator
could be a battery bank with three-phase inverter or a fossil-fired generator
set. The speed of the other generator would be regulated according to a
fixed schedule proportional to DNI. The dish Stirling with induction generator would then follow the other generator and together they would provide
power to the water pump. The use of a DC generator and DC water pump
motor, or a DC generator with three-phase inverter was also considered.
Photovoltaic field experience indicates that there are significant reliability
concerns with DC water pumps and pump controllers.
The power management selected for the Sandia stand-alone dish Stirling
system utilizes a three-phase synchronous generator to directly drive a
10-HP, three-phase, 480 V induction submersible water pump motor. This
approach was the simplest, lowest cost, and highest efficiency option considered, but also had technical risk. The synchronous generator is selfexcited and incorporates a microprocessor-controlled field to produce
voltage that is proportional to speed, and speed is proportional to power.
This approach has the advantage of utilizing standard water-pumping hardware and opens up possibilities for driving other single-motor applications
that can accommodate a variable rotational speed.
Fig. 9.28 shows a plot of key operational parameters during a typical
startup of a remote dish Stirling system. During startup, engine speed is
regulated by the engine control system at about 800 rpm. At this speed the
generator does not generate power and the instantaneous loads associated
with starting the induction motor are avoided. After the engine warms up
and is capable of producing power without stalling, the power conversion
unit controls allow the engine speed to gradually increase and the load of
the motor to be gradually introduced. The engine then operates, for example,
at a speed proportional to insolation level, typically between about 1400
and 1,850 rpm. Note the idle mode at around 800 rpm, where no power is
generated. When the engine is released from idle, the generator begins to
© Woodhead Publishing Limited, 2012
Concentrating solar power technology
800
100
80
400
Voltage
60
Water flow
200
40
1000
Max receiver
Engine speed (RPM)
600
Water flow (l/min)
Temperature (°C), voltage
120
Engine
speed
Heater
head
1500
140
500
312
0
0
08:59
09:00
Time (MST)
0
20
09:01
9.28 Typical startup of a dish Stirling off-grid system.
generate power. The heater head temperature is indicative of the warm-up
of key engine components, which occurs much more slowly than the lowthermal-mass receiver tubes.
9.6
Optimization of manufacture
9.6.1
Reflector fabrication
Facetted paraboloids (i.e., consisting of individual spherical or parabolically
shaped segments) and full-surface paraboloids are encountered in different
design approaches. For facetted concentrators, several mirror segments are
mounted on a supporting structure. The segments are attached and oriented
individually. Such mirror segments may either consist of glass mirrors or
substrates covered with reflecting foil or thin-glass mirrors.
For full-surface concentrators, the entire concentrator surface is shaped
parabolically by a forming process. For instance, the spb membrane dishes
use a pre-stressed metallic membrane attached on both sides to a stable
ring (stretched membrane technology). Subsequently, it is transformed into
the desired shape via a forming process (e.g., by water load) and stabilized
via a controlled vacuum. Such low weight metal membrane designs provide
full-surface concentrators with high rigidity and high optical quality. Alternatively, the surface may be made from sandwich elements made of fibre
glass reinforced epoxy resin with thin glass mirrors glued onto them (e.g.,
the EuroDish). For the small dishes employed by Infinia, injection moulded
plastic elements are used.
© Woodhead Publishing Limited, 2012
Parabolic dish concentrating solar power (CSP) systems
313
With many dish concentrators, the reflector is made up of facets mounted
on a supporting structure. There is a considerable variety of facet designs
that have been realized:
•
Sandwich substrates from reinforced plastic or metal (e.g., WG Associates/Cummins WG-500 dish) with bonded thin glass mirrors (typically
0.8–1.2 mm)
• stamped steel (e.g., SES Suncatcher) or composite plastic substrates
(e.g., Infinia Power dish) with thin glass mirrors
• above substrates can also be equipped with metallized plastic foils or
anodized aluminium sheets
• self-carrying slumped glass mirrors are the standard solution for parabolic trough concentrators but have not been applied for dishes except
single prototypes.
With the quantities of dish concentrators likely to increase, facet manufacturing technologies for large-scale production like stamping have become
more important. They require major investments in tooling but can achieve
low cost and high precision. An important field of optimization and cost
reduction for thin glass mirrors is bonding onto the substrate. The initially
flat thin mirror tiles are limited in size to withstand the biaxial bending and
handling issues, thus a considerable number of tiles may be required for
every facet. Robotic application with adhesive foils or sprayed fluid glue
can be used, but is nevertheless a significant cost factor. Metallized plastic
foils and anodized aluminium could achieve a cost reduction in this process.
However, verification of their long-term stability in outdoor applications is
still ongoing, therefore they have not yet been used in major projects.
The support structures are frameworks or truss/girder systems from steel
sections in most cases. Typically, subassemblies are welded and corrosion
protected in the factory and bolted together on the erection site. Parabolic
trough collectors have proven that this concept can be cost effective even
in large quantities. Alternatively, smaller dishes have also been made up
from stamped parts, and this has the potential for further cost reduction as
stamped parts have low specific cost. Since the maximum size of stamped
parts is limited by the availability of large presses, this procedure has not
been applied for bigger concentrators.
9.6.2
Drives
Many pylon-based dish designs use slewing drives for azimuth and linear
drives for elevation movement. Both azimuth and elevation drives can
profit from the large market for tracking PV and heliostats where similar
solutions are used, and suppliers are already manufacturing large volumes.
One difference from PV applications is that high stiffness and low backlash
© Woodhead Publishing Limited, 2012
314
Concentrating solar power technology
are a necessity. However, slewing drives in particular are a considerable cost
factor which calls for new and advanced solutions.
Turntable-based dish mounting and tracking (e.g., EuroDish) is characterized by small and cost-effective gears but needs many structural parts to
be assembled and large foundations, for big systems to be equipped with
rails. Their application is mainly for large dishes (e.g., ANU dishes) where
the cost for pylon-based drives is too high and their stiffness is too low.
9.6.3
Trade-off between concentrator accuracy and cost
The optical quality of the concentrator is the major factor for concentrator
performance. If power conversion units with high operating temperatures
like Stirling engines are to achieve reasonable efficiency, they need a
compact absorber or, in the case of a cavity receiver, a small aperture to
reduce infrared radiation and convection losses. For this, a dish with high
optical quality and a high concentration factor is required, which is directly
coupled to the need for high accuracy of the reflective surface. Besides the
desire for a highly specular reflection, a characteristic of solar concentrator
optics is that slope deviations from nominal are the most relevant factor.
For high performance dish concentrators, the average slope errors typically
range between 0.5 and 3 mrad (0.03 to 0.17°). High precision is required to
achieve numbers towards the lower end, which is highly cost related and
therefore an important factor for the economic analysis.
Depending on the design and shape of the absorber, the alignment of
facets may have to be tailored to achieve a flux distribution adapted to the
needs of the power conversion unit.
The optical quality impact on concentrator performance and the allowable cost for reflectors with different surface slope errors were studied in
Andraka (2008). The outcome was that high precision reflectors pay out in
the end even at relatively high cost.
Other factors for concentrator performance are stiffness of the structure
under dead weight and wind load, facet alignment in the case of multifacet
concentrators, alignment of the absorber along and across the optical axis
and tracking accuracy when following the sun.
Figures 9.29–9.35 show how the above-mentioned errors and deformations affect the reflector contour and the flux distribution on the absorber
of a Stirling engine. A dish consisting of six gore-shaped facets was used for
this example.
A comprehensive system optimization has to consider all the above
factors. Since the optimum is always cost driven, achieving an economic
concentrator design needs detailed knowledge of the cost of different manufacturing and assembly methods.
© Woodhead Publishing Limited, 2012
315
40
30
0.2
0.15
0.1
0.05
0
–0.05
–0.1
–0.15
–0.2
20
10
0
–0.2
–0.1
z
Flux (W/cm2)
Parabolic dish concentrating solar power (CSP) systems
1
0.8
0.6
0.4
0.2
0
–0.2
–0.4
–0.6
–0.8
–1
–2.5
y (m)
0
0.1
2.5
2
1.5
1
0.5
–2 –1.5
–1 –0.5
0
0.5
0.2
x (m)
1
0
–0.5
–1
–1.5
–2
–2.5
y
1.5
2
2.5
x
60
50
40
30
20
10
0
–0.2
×10–3
4
2
0
–2
–4
–6
–2.5 –2
0.2
0.15
0.1
0.05
0
–0.05
–0.1
–0.15
–0.2 y (m)
2.5
2
1.5
1
0.5
z
Flux (W/cm2)
9.29 Ideal undeformed structure.
–0.1
0
0.1
x (m)
0.2
–1.5
–1
–0.5
0
0.5
1
x
1.5
2
2.5
0
–0.5
–1
–1.5
–2
y
–2.5
9.30 Structure under dead weight and wind.
2
1.5
0.2
0.15
0.1
0.05
0
–0.05
–0.1
–0.15
–0.2 y (m)
×10–3
2
1
0
–1
–2.5 –2
1
0.5
z
Flux (W/cm2)
2.5
60
50
40
30
20
10
0
–0.2
–0.1
0
0.1
x (m)
0.2
–1.5
–1
–0.5
0
0.5
1
x
1.5
9.31 Isolated effect of reflector element waviness.
© Woodhead Publishing Limited, 2012
0
–0.5
–1
–1.5
–2
–2.5
y
2
2.5
316
Concentrating solar power technology
1.5
1
40
0.2
0.15
×10–3
0.1
1
0.05
0
–1
0
–2.5 –2
–0.05
–0.1
–0.15
y (m)
–0.2
30
20
10
0.5
0
–0.5
z
Flux (W/cm2)
2.5
2
50
0
–0.2
–0.1
0
0.1
–1
–1.5
–1.5
–1
–0.5
0
0.5
0.2
x (m)
1
y
–2
1.5
2
2.5
x
–2.5
2.5
2
1.5
50
40
0.2
×10–3
0.15
2
0.1
1
0.05
0
0
–1
–0.05
–2.5 –2
–0.1
–0.15
y (m)
–0.2
30
20
1
0.5
z
Flux (W/cm2)
9.32 Isolated effect of reflector element support point deviations.
10
0
–0.2
–0.1
0
0.1
x (m)
–1.5
–1
–0.5
0
0.5
0.2
1
y
0
–0.5
–1
–1.5
–2
–2.5
x
1.5
2
2.5
40
30
0.2
0.15
0.1
0.05
20
10
0
–0.2
–0.1
y (m)
0
0.1
x (m)
0
–0.05
–0.1
–0.15
–0.2
–0.25
z
Flux (W/cm2)
9.33 Isolated effect of reflector element tilt.
0.2
1
0.8
0.6
0.4
0.2
0
–0.2
–0.4
–0.6
–0.8
–1
–2.5
2.5
2
1.5
–2 –1.5
–1 –0.5
0
0.5
1
1.5
2
2.5
x
9.34 Isolated effect of target misalignment from optical axis.
© Woodhead Publishing Limited, 2012
1
0.5
0
–0.5
–1
–1.5
–2
y
–2.5
70
60
50
40
30
20
10
0
–0.2
317
2.5
0.2
0.15
0.1
0.05
0
–0.05
–0.1
–0.15
–0.2
z
Flux (W/cm2)
Parabolic dish concentrating solar power (CSP) systems
–0.1
0
0.1
x (m)
0.2
y (m)
×10–3
6
4
2
0
–2
–4
–6
–8
–2.5 –2
2
1.5
–1.5
–1
–0.5
0
0.5
1
1.5
x
1
0.5
0
–0.5
–1
–1.5
–2
–2.5
y
2
2.5
9.35 All errors plus dead weight and wind.
9.6.4
Strategies for site assembly and alignment
The cost of site assembly and alignment is an important fraction of the total
system cost, thus it is important to find an efficient solution. The number of
units to be erected on a site plays a prominent role in the decision on
assembly strategy. Large installations allow for extensive tooling and
machinery, while small clusters or even single system erection call for leaner
methods.
For medium to large plants, a site workshop is erected as is common
practice with parabolic trough collectors.
Automation is of course a way to reduce required manpower; the automotive industry is the oft-cited standard. However, the conditions for use
of robots, for example, are different in a site assembly workshop: it usually
runs for a relatively short period which increases the share of run-in, the
environmental conditions like temperature changes and soiling are harder
to create and the availability of personnel and spare parts for correction of
failures is worse. Therefore, deployment of automated processes has to be
well planned.
For one of the most critical assembly steps, joining and aligning of reflector elements, at least the following principles can be differentiated:
• use of precise jigs for the concentrator structure assembly ensures that
the mirror supports are exactly in place,
• mirrors placed on a precision jig, imprecise support structure attached
to the mirrors via a tolerance compensating joining method,
• building the structure from precise parts with low play in the joints
results in a precise assembly and mirror supports at the right
positions,
• less precise structure with adjustable mirrors and a suitable alignment
procedure (Andraka et al., 2010).
© Woodhead Publishing Limited, 2012
318
Concentrating solar power technology
Also, transportation of dish assemblies from the site workshop into the
installation field and mounting on the prepared foundation or pylon have
to be organized efficiently. Trucks or special vehicles and mobile cranes are
employed for this purpose.
Finally, fast and standardized commissioning procedures with little personnel demand are required. For large plants, commissioning can be conducted in sections according to the construction progress, making use of the
modular character of decentralized power generation. Thus parts of the
plant can already produce energy during plant construction and reduce
interim financing.
9.7
Future trends
9.7.1
Decentralized applications
For a long time, parabolic dish and dish Stirling systems have been regarded
mainly as an option for decentralized applications in the range of some kW
to some MW. However, most of the product developments in recent years
focus on large installations from 10 MW to almost 1 GW. The reasons are,
on the one hand, that cost prognosis leads to the expectation that competitiveness with parabolic trough and PV can be achieved and that the market
in many countries requires bulk solar power. On the other hand, operational experience with dish Stirling technology spread over more than 25
years has been collected with prototypes usually run under test conditions.
Commercial applications are different in many aspects, thus lessons learned
from prototypes cannot be directly transferred. Furthermore, the number
of units under test has always been small. The number of operational hours
collected so far has not been sufficient for maturing such a complex system
and for achieving the targeted reliability and low maintenance and repair.
Since the business plans and the market demands do not allow for extensive
further testing in small and medium size installations, teething troubles have
to be expected. Resolving future issues and implementation of necessary
retrofitting would be very costly in decentralized applications. Large installations allow for better supervision of the systems and reduce the costs of
unplanned maintenance and repair.
Dish Stirling systems are suitable for decentralized applications but it can
be expected that this market will be served at a later stage when the technology has matured and has proven reliability and low maintenance costs.
9.7.2
System size
Most of the present dish Stirling product developments range between
some kW and 25 kW. Going for small units increases relative cost for
© Woodhead Publishing Limited, 2012
Parabolic dish concentrating solar power (CSP) systems
319
components like electrics, controls and field cabling. For the same installed
power, more parts have to be assembled. On the other hand, small size
raises the number of produced parts and may enable thresholds in production volume to be reached for large mass production, especially for manufacturing technologies known from the automotive industry. The Infinia
3 kW system is an example of this path, where stamped parts are used for
much of the frame. Also, wind moments on small dishes and thus loads for
the drive system reduce more than proportionally with decreasing dish
surface area. Wind moments also affect the dish structure and lead to
reduced specific mass (kg/m2 of projected concentrator area). In contrast,
large systems have the benefit of lower part count to be assembled but they
suffer from increasing specific mass.
The economic optimum in size depends on many variables. Many designers found that there is an optimum size in the range of 10–25 kWel, corresponding to dish projected areas of 50–120 m2. However, today we see
commercial and pre-commercial dish and dish Stirling systems in very different sizes. The upper end is marked by the ANU Big Dish with 400 and
500 m2; HelioFocus of Israel is also developing a very large dish. These
systems have solar-chemical reactors or receivers that provide steam. The
lower end is marked by the Infinia dish Stirling system with 3 kW and 15 m2.
Especially for dish Stirling, a major design driver for selecting the system
size is always the availability of suitable engines which sometimes overrides
what has been found to be the theoretical optimum.
9.7.3
Energy storage
A major advantage of centralized solar thermal power plants as parabolic
trough and power tower systems is the possibility to add thermal storage.
The storage option allows for increasing the capacity factor which is an
important criterion in an electrical power distribution system with limited
transmission line capacities and without sufficient electrical storage. It is
even more important for the dish Stirling technology since their thermal
inertia is low and electrical power output directly follows the solar radiation. Developing thermal storage for dish engines would therefore be very
helpful to gain share in future markets.
There are two options for a thermal storage: it can be located at the
engine itself or it might be a central unit in a power plant. A small local
storage has the issue that it increases the engine dead weight. Since with
today’s storage technologies a useful storage capacity of at least some hours
at nominal load means a multiple of the engine mass, this requires adding
substantial extra material into the engine support structure, thus increasing
loads on dish drives and bearings. Furthermore, small storage has an unfavourable surface-to-volume ratio and thus high thermal losses from thermal
© Woodhead Publishing Limited, 2012
320
Concentrating solar power technology
conduction through the insulation result. A central storage solution faces
the issue of transporting high temperature heat over considerable distances
and across movable joints to accommodate the dish movement. Keeping
thermal losses over all the lines to the storage low is a challenge.
Both storage options require a high temperature level to supply the
engines; with typical Stirling engine working gas temperatures between 600
and 700°C, the storage temperature needs to be even higher to overcome
required temperature differences for the heat transfers. So far no dish Stirling system with thermal storage has been published, which also indicates
the sizeable technical and economic difficulties. Therefore the storage
option is attractive, but there are still doubts whether viable technical and
economic solutions can be found.
9.7.4
Hybrid operation
Similar to plants with thermal storage, hybrid solar power plants allow for
de-coupling solar incident power and electric energy generation. In many
cases, they can benefit from better feed-in tariffs, i.e. by avoiding losses of
increased rates during demand peak times. The feasibility of hybridizing
dish Stirling units has been proven in some projects (Laing and Reusch,
1998; Moreno et al., 1999), but to date hybrid dish Stirling systems have not
left the experimental stage. Perhaps the most challenging task is to find a
technical solution for the conflictive demands of solar receivers and fossil
or biomass-fired heaters at reasonable cost. With known solutions, the loss
in efficiency compared to solar only systems and/or complexity and cost are
too high.
If an efficient and reliable hybrid system at reasonable cost can be developed in the future, this could open considerable market opportunities.
Today, it seems that this will not be achieved in the short term.
9.8
Conclusion
The parabolic dish concentrator development over the last 25 years has
demonstrated an impressive diversity of designs and solutions. While many
systems apply a Stirling engine as a PCU, others generate heat to be supplied by other thermodynamic cycles. Two-axis solar concentration allows
for high upper process temperatures and the highest conversion efficiencies
of all solar concentrating technologies. On the other hand, the reflector
geometry demands more in terms of manufacturing and the bi-axial tracking to the sun requires additional effort. There is still a considerable variety
of designs and a large span of concentrator sizes with present developments.
First commercial products are going to enter the market and will have to
prove their technical maturity and economic viability. Thermal storage and
© Woodhead Publishing Limited, 2012
Parabolic dish concentrating solar power (CSP) systems
321
hybrid operation are attractive options to improve economic viability but
are technically challenging and not yet proven.
9.9
Sources of further information and advice
Mancini, T., Heller, P., Butler, B., Osborn, B., Schiel, W., Goldberg, V., Buck,
R., Diver, R., Andraka, C. and Moreno, J. (2003) ‘Dish-Stirling systems:
an overview of development and status’, Journal of Solar Energy
Engineering, 125 (2), 135–151.
Stine, W.B. and Diver, R.B. (1994) ‘A compendium of solar dish/Stirling
technology’, Report SAND93-7026. Sandia National Laboratories,
Albuquerque, NM.
9.10
References and further reading
Andraka, C.E. (2008) ‘Cost/performance tradeoffs for reflectors used in solar concentrating dish’, Proc. 2nd International Conference on Energy Sustainability,
August 10–14, Jacksonville, FL.
Andraka, C., Diver, R., Adkins, D., Rawlinson, S., Cordeiro, P., Dudley, V. and Moss,
T. (1993) ‘Testing of Stirling Engine Solar Reflux Heat-Pipe Receivers’, Proc. 28th
Intersociety Energy Conversion Conf. (IECEC), Atlanta, GA, August.
Andraka, C., Adkins, D., Moss, T., Cole, H. and Andreas, N. (1995) ‘Felt-Metal-Wick
Heat-Pipe Receiver’, Solar Engineering 1995, Proc. ASME/JSME/JSES Int. Solar
Energy Conf., Maui, HI, March.
Andraka, C., Yellowhair, J., Trapeznikov, K., Carlson, J., Myer, B. and Stone, B. (2010)
‘AIMFAST: An Alignment Tool Based on Fringe Reflection Methods Applied to
Dish Concentrators’, Proc. SolarPACES 2010 Conference, Sepember 21–24,
Perpignan, France.
Bean, J.R. and Diver, R.B. (1995) ‘Technical Status of the Dish Stirling Joint Venture
Program’, Proc. 30th IECEC, Orlando, FL, pp. 2.497–2.504.
Buck, R., Heller, P. and Koch, H. (1996) ‘Receiver Development for a Dish-Brayton
System’, Proc. ASME Int. Solar Energy Conference, San Antonio, TX, April.
Diver, R.B., Andraka, C., Rawlinson, K., Moss, T.A., Goldberg, V. and Thomas, G.
(2003) ‘Status of the Advanced Dish Development System Project’, ASME 2003
International Solar Energy Conference (ISEC2003), March 15–18, Kohala Coast,
Hawaii.
EPRI Report (1986) ‘Performance of the Vanguard Solar Dish Stirling Engine
Module’, Electric Power Research Institute, AP 4608, Project 2003-5.
Gallup, D. and Mancini, T. (1994) ‘The Utility-Scale Joint-Venture Program’, Proc.
29th IECEC, August 7–12, Monterey, CA.
Gordon, J. (2001) Solar Energy, The State of the Art, ISES Position Papers, James &
James, London.
Jaffe, L.D. (1988) ‘A review of test results on solar thermal power modules with
dish-mounted Stirling and Brayton cycle engines’, Journal of Solar Energy Engineering, 110, 275–281.
Kaneff, S. (1991) The White Cliffs Project – Overview for the period 1979–89. NSW
Office of Energy, Sydney, Australia.
© Woodhead Publishing Limited, 2012
322
Concentrating solar power technology
Koshaim, B. (1986) ‘Report: Fifty KW Solar Membrane Concentrator’, The
SOLERAS Program, Saudi Arabian National Center for Science and
Technology.
Laing, D. and Reusch, M. (1998) ‘Design and Test Results of First and Second generation Hybrid Sodium Heat Pipe Receivers for Dish Stirling Systems’, Proc.
ASME International Solar Energy Conference, Albuquerque, NM.
Lopez, C. and Stone, K. (1992) ‘Design and Performance of the Southern California
Edison Stirling Dish’, Solar Engineering, Proc. ASME Int. Solar Energy Conf.,
Maui, HI, April 5–9.
Lopez, C. and Stone, K. (1993) ‘Performance of the Southern California Edison
Company Stirling Dish’, SAND93-7098, Sandia National Laboratories, Albuquerque, NM.
Lovegrove, K., Burgess, G. and Pye, J. (2011) ‘A new 500 m2 paraboloidal dish solar
concentrator’, Solar Energy, 85 (4), 620–626.
Mancini, T., Heller, P., Butler, B., Osborn, B., Schiel, W., Goldberg, V., Buck, R., Diver,
R., Andraka, C. and Moreno, J. (2003) ‘Dish-Stirling systems: an overview of
development and status’, Journal of Solar Energy Engineering, 125 (2), 135–151.
Moreno, J., Rawlinson, S., Andraka, C., Mehos, M., Bohn, M.S. and Corey, J. (1999)
‘Dish/Stirling Hybrid receiver Sub-Scale Tests and Full-Scale Design’, 34th Intersociety Energy Conversion Conference, Vancouver.
SouthWest Solar (2011) http://www.swsolartech.com/ accessed December 2011.
Stine, W. (1993) ‘An International Survey of Parabolic Dish Stirling Engine Electrical Power Generation Technology’, Solar Engineering, Proc. ASME/ASES Joint
Solar Energy Conf., Washington, D.C.
Stine, W.B. and Harrigan, R.W. (1985) Solar Energy Fundamentals and Design: with
computer applications. John Wiley & Sons, New York.
Stine, W.B. and Diver, R.B. (1994) ‘A compendium of solar dish/Stirling technology’,
Report SAND93-7026. Sandia National Laboratories, Albuquerque, NM.
© Woodhead Publishing Limited, 2012
10
Concentrating photovoltaic (CPV) systems
and applications
S. H O R N E, SolFocus Inc., USA
Abstract: This chapter is a summary of the state of the art of
concentrating photovoltaic (CPV) systems, discussed from several
viewpoints. It begins with an abbreviated history of the technology, then
continues to a discussion of the characteristics, market, and system
design. A short piece on future trends concludes. Because of its complex
nature, this chapter is limited to a qualitative introduction to this
interesting and growing field, and assumes a general familiarity with
photovoltaics.
Key words: concentrating photovoltaic systems (CPV), multi-junction
photovoltaic cells, optical systems, acceptance angle, spectral transfer
function, two-axis tracking, levelized cost of energy, high volume
manufacturing.
10.1
Introduction
Concentrating photovoltaic (CPV) systems operate by using an optical
assembly to concentrate light onto a photovoltaic (PV) cell. In other words,
they entrain a large area of solar energy onto a small cell, which operates
at an irradiation level many times greater than that of direct, unconcentrated sunlight. A PV cell’s conversion efficiency actually improves somewhat with increasing irradiation levels (Olson et al., 2007), and will deliver
much more power when used under concentration than when operated
under direct sunlight. CPV technology exploits this to significantly reduce
the cost of energy by amortizing the cost of the cell and attendant optics,
housings and tracking systems over the high energy output. In practice,
concentration ratios (while there are several definitions, broadly the ratio
of the irradiance on the cell to the irradiance at the entrance aperture of
the concentrator), generally expressed in ‘suns’, fall into two general groupings. Low concentration photovoltaic (LCPV) devices operate between 1.25
and approximately 40 suns, and high concentration photovoltaic (HCPV)
devices have been built between 250 and 1,700 suns. Devices with concentration levels between HCPV and LCPV (medium concentration photovoltaic devices or MCPV) have not received much attention, mainly due to
their economics, as will be illustrated later.
323
© Woodhead Publishing Limited, 2012
324
Concentrating solar power technology
CPV can be applied with any of the CSP concentrator system types that
have been discussed in other chapters (trough, tower, linear, Fresnel or
dish). The concentrators are, however, designed specifically for the CPV
application. Fresnel lens approaches are also applied to CPV.
This chapter introduces CPV systems from several different viewpoints,
and describes their role in the solar marketplace.
10.1.1
Historical summary
While CPV technology has been under development for many years, commercialization has been elusive as technical and reliability difficulties dominated the development of this seemingly simple idea (Rosenthal and Lane,
1991). In addition, the rapidly maturing silicon panel market with its head
start of many decades raised significant barriers to market entry. Swanson
(2002: 449–452) illuminates the dilemma facing the CPV industry in greater
detail than is possible here, and explains why the expected increase in commercial investment did not occur during that period. Sala and Luque (2007:
1–11) characterize the period up to the late 1990s as one dominated by
academic leadership, with some product development progress being made
by only a small number of companies, Arco Solar (Rosenthal and Lane,
1991), Amonix and Solar Systems Australia being three examples.
Within the past ten years, though, advances in the efficiency of practical
high-performance multi-junction cells, first described by Olson and Kurtz
(1993) from the US National Renewable Energy Labs (NREL), have reignited interest in HCPV. These cells, first developed for the space applications market and using materials other than silicon, promise conversion
efficiencies of well over 40% but at a cost extremely prohibitive for use in
standard panels, at one sun. The only possible application for these cells in
a terrestrial environment is in a HCPV system. Significantly, this rebirth
occurred at a time of great interest in energy prices and sustainable practices, much of it coming from the worldwide venture capital community,
seeking a post-internet boom market. In addition, by the time the investment industry started to seriously analyze CPV, these cells had passed the
stringent reliability standards of the space industry, and had amassed millions of successful cell-hours of operation. This confluence of performance
promise, interest in renewable energy sources, and positive reliability data
emboldened the investment industry, and an explosion in new HCPV companies occurred in the first few years of the century. This was soon followed
by investments into the cell segment itself, as meaningful progress had been
made on new cell morphologies and related technologies (Miyashita et al.,
2007), building on the pioneering work at NREL.
The LCPV segment also received significant interest. Improvements in
silicon cells, while not as spectacular as those in multi-junction cells, were
© Woodhead Publishing Limited, 2012
Concentrating photovoltaic (CPV) systems and applications
325
important to this segment of the industry. The combination of efficiencies
in the 18% range and low fabrication costs allowed for designs that made
economic sense. Though a smaller segment than HCPV, LCPV has attracted
high-quality commercial representation.
While all concentrator optics are constrained by the physics of reflection,
refraction and total internal reflection (TIR), within these limits the relatively new field of non-imaging optics (NIO) pioneered by Welford and
Winston (1989) added opportunity for innovation. Significant performance
and manufacturability improvements have been realized by applying NIO,
and the pathway to practical, deliverable products has become much more
navigable. As a result, many new CPV companies have worked to merge
NIO with the new cells, and a large range of designs have recently appeared
at both ends of the concentration range, further contributing to what was
an already well prototyped field.
Today, the leading companies in CPV have matured their products, have
commissioned high-volume production lines and have amassed large
amounts of data from operating installations. The focus for many of these
companies is now on proving their bankability and product reliability, as
larger commercial opportunities become available.
Currently, there are over 20 active CPV companies, and Table 10.1 summarizes the status of the leaders. Note that a more detailed description of
the techniques used by many of these companies appears later in this
chapter, as does a sample of product photographs.
After a long gestation, CPV is starting to meet its promise. Swanson
(2002: 449–452) declared that CPV is ‘a long range option of vital importance to the energy security of the world. Cost analyses indicate that it
certainly has the possibility of becoming the low-cost PV approach in large
installations.’ That long-range timeframe is upon us.
10.2
Fundamental characteristics of concentrating
photovoltaic (CPV) systems
To understand the appeal of the technology and the contribution it can
make to the already large offering of solar technologies, it is necessary to
understand its characteristics. Some central concepts are discussed in this
section.
10.2.1
Acceptance angle
Regardless of the concentration level or optical method used, a CPV system
can be thought of as telescope placed in front of an efficient PV cell. Only
light entering this telescope will reach the cell and be converted to electricity, which presents a limitation not seen in standard flat plate or one-sun
© Woodhead Publishing Limited, 2012
326
Concentrating solar power technology
Table 10.1 Photovoltaic concentrator manufacturers in or close to production
as of August 2010
Company
System
Installations
(completed or
under construction;
not announced)
SolFocus
High concentration, multi-junction cell,
glass, point-focus reflective optical
array
High concentration, silicon cell, acrylic,
point-focus refractive (Fresnel lens)
array
High concentration, multi-junction cell
acrylic, point-focus refractive (Fresnel
lens) array
Medium concentration, silicon cell,
single acrylic line focus refractive
(Fresnel lens) module
High concentration, multi-junction cell,
glass/silicone point-focus refractive
(Fresnel lens) array
High concentration, multi-junction cell,
point-focus refractive (Fresnel lens)
array
High concentration, multi-junction cell,
large single piece glass point-focus
reflector (parabolic) system
High concentration, multi-junction cell
glass/silicone, point-focus refractive
(Fresnel lens) array
Low concentration, silicon cell, large
single piece aluminum line focus
reflective system (DSMTS system,
described later)
8 MW
Amonix
Amonix
Entech
Concentrix
Opel
Solar
Systems/
Silex
Emcore
Skyline Solar
Approx. 9 MW
0.38 MkW
Approx. 0.2 MW
1.7 MW
0.33 MW
Approx. 1 MW
1.0 MW
0.1 kW
systems. Because of this, a concentrator system possesses a field of view or
acceptance angle, which is inextricably linked to the concentration ratio. For
the general rotationally symmetric concentrator, the relationship for the
maximum achievable geometric concentration ratio, as previously discussed
in Chapter 2, is derived by Welford and Winston (1989: 27) to be:
⎛ n sin φ ⎞
C g = ⎜ out
⎝ nin sin θ ⎟⎠
2
[10.1]
where Cg is the maximum possible concentration ratio, nout is the index of
refraction of the output medium of the concentrator, φ is the half-angle
of the edge ray emerging from the concentrator output, nin is the index of
© Woodhead Publishing Limited, 2012
Concentrating photovoltaic (CPV) systems and applications
327
refraction of the input medium of the concentrator, and θ is the half-angle
of the edge ray entering the concentrator: the acceptance angle.
Rearranging this equation for acceptance angle gives:
⎛ 1
θ = sin −1 ⎜
⎝ nin
(nout sin φ)2 ⎞
⎟
Cg
⎠
[10.2]
From inspection, the maximum theoretically attainable acceptance angle
would occur for φ = 90°, when the output rays are allowed to emanate over
the entire hemisphere from the concentrator’s exit aperture. A concentrator
of 1,000 suns, operating in air (n = 1) with a final concentration stage constructed of low-iron glass (nout = 1.5, approximately) would have an acceptance angle of ±2.7°. However, since PV cells exhibit an approximate cosine
relationship between input angle and conversion efficiency (Spectrolab,
2010), in practice this number is not attainable, as the high angle light would
not be efficiently utilized. If the largest admissible output half-angle is taken
to be at the point at which the cell output power is 90%, or ±26°, the acceptance angle is reduced to ±1.2°.
Because of this, CPV systems are mostly sensitive to the direct component of the sun’s radiation only – that which emanates directly from the sun
and is not reflected, refracted or scattered by the atmosphere or terrestrial
objects. The term used to describe the direct or beam power is direct irradiance. In practice, as a design approaches the workable limit for acceptance
angle, it becomes more expensive and, possibly, less efficient. Central to all
concentrator design is the interesting and difficult challenge of the four-way
trade between efficiency, cost, manufacturability, and acceptance angle.
10.2.2
Principles of photovoltaic devices
CPV systems convert light to electricity through the use of a photovoltaic
cell, and their electrical characteristics run parallel to those of standard, flat
silicon solar panels. While an adequate treatment of the physics of photovoltaics is far outside the scope of this chapter, an introduction to the topic
is warranted. For a complete development, Green (1998) is very concise.
Photovoltaic devices comprise semiconductor materials that convert light
to electricity in a very direct manner. These devices are made from crystalline materials deliberately ‘doped’ with impurities that donate additional
weakly bonded electrons to the crystal. When in their unexcited state, these
electrons occupy a range (or band) of energies called the valence band. If
one of these electrons gains sufficient energy from, for example, an interaction with a photon, it enters a higher energy band called the conduction
band. Energies in the conduction band exceed those allowed in the valence
band by an amount defined by the bandgap of the crystalline material. The
bandgap represents a forbidden energy zone, over which the electrons
© Woodhead Publishing Limited, 2012
328
Concentrating solar power technology
effectively jump. The bandgap is primarily a function of the materials (e.g.,
silicon, gallium arsenide, and germanium to name a few) and temperature.
Bandgaps are expressed in electron volts (eV), the amount of energy
needed to move one electron through a potential of one volt. In order to
excite an electron from the valence to the conduction band, a photon must
have energy at least equal to the bandgap. Photon energy in excess of the
bandgap is wasted as heat. Once in the conduction band, excited electrons
are highly mobile, and can be directed out of the material to an electrode,
forming a DC electric current.
Crucial to the extraction of the electrons from the crystal lattice is the
structure formed by combining two types of semiconductors, namely n-type
and p-type, into a diode arrangement, commonly called a photodiode when
used for electricity generation. Without this, any freed electrons would
quickly fall back into the valence band, giving up their energy as heat or
radiation via a multitude of complex processes grouped under the general
term recombination. Recombination exists right through the crystal lattice,
but at the surface is particularly strong. The design of the diode and
its manufacturing processes can minimize but not fully eliminate
recombination.
Photons have energies that are inversely proportional to their wavelength. Photons of blue light (wavelengths on the order of 400 nm to
470 nm) are more energetic than photons of red (620 nm to 750 nm), and
are generally absorbed closer to the surface of a semiconductor. Photons
with energies greater than the bandgap will be able to produce an electric
current, the surfeit energy they contain being given up thermally and lost.
Because of this, and also because highly energetic photons are more subject
to surface recombination, as photon energy increases – or wavelength
decreases – from the minimum value required to activate a bound electron,
device conversion efficiency decreases.
Similarly, other photons within the solar spectrum having energy less than
the bandgap energy cannot promote an electron from the valence band to
the conduction band. Instead, these photons pass through the material,
eventually to be adsorbed in large fraction and converted to waste heat via
various unwanted processes inherent in semiconductor materials.
Taken together, the above two mechanisms will cause the device’s conversion efficiency to peak at a particular wavelength, and to exhibit sensitivity within a specific range of wavelengths only, dependent on the characteristic
bandgap energy. Kurtz and Geisz (2010: A75) describe the optimized
bandgap for the solar spectrum to be approximately 1.4 eV. Monocrystalline silicon (c-Si) has a bandgap of 1.11 eV, which is very close, and is one
reason why silicon has dominated the photovoltaic industry to date. The
characteristics of silicon can be seen in Fig. 10.1, where the efficiency is
plotted against the wavelength of light. Also shown on the graph is the
© Woodhead Publishing Limited, 2012
Concentrating photovoltaic (CPV) systems and applications
329
120%
1.80
Spectral density (W/m2-nm)
100%
1.40
80%
1.20
1.00
60%
0.80
40%
0.60
0.40
20%
Relative response of Si photocell
1.60
0.20
0%
4000
5000
AMI 1.5 spectrum
Si panel relative response
0.00
0
1000
2000
3000
Wavelength (nm)
10.1 AMI 1.5 spectrum, overlaid with silicon junction conversion
relative efficiency plot.
Rs
+
Rp
−
Iph
(α Irradiation)
10.2 Single-junction solar cell, equivalent circuit.
spectrum of the sun under AMI 1.5 conditions. This is a standard spectrum,
and is the result of the modification of the sun’s native spectrum by 1.5
times the standard atmospheric depth, which occurs with the sun 48° above
the horizon (chosen as a reasonable average estimate for the entire day).
As can be seen, silicon’s characteristics limit to converting a portion, but
not all, of the sun’s incident energy to electricity, leading to the upper bound
efficiency. Green (1998: 89–90) shows this limit to be approximately 27%.
At the device level, the photodiode can be modeled by the equivalent
circuit of Fig. 10.2, with the relationship between voltage and current shown
in the solid curve in Fig. 10.3 (commonly called an IV curve). The product
of the current and voltage produces a power–voltage curve, shown as a
© Woodhead Publishing Limited, 2012
330
Concentrating solar power technology
120%
Output current (% of max)
Output power (% of max)
Short circuit current (Isc)
Maximum power point:
Voltage - Vmp
Current - Imp
100%
80%
60%
40%
20%
Open circuit voltage (Voc)
0%
0%
20%
40%
Output voltage (% of max)
60%
80%
IV curve
100%
120%
Power
10.3 IV and power curves, photovoltaic junction under illumination.
broken curve in this figure, and which illustrates one important aspect of
the application of these devices: they need to operate at the correct voltage,
the maximum power point voltage (or Vmp), or they will not produce their
maximum output. This maximum power point is continuously searched for
by photovoltaic power management systems, for example inverters.
Parasitic series and parallel resistances (Rs and Rp in Fig. 10.2) will cause
departure from ideal photodiode characteristics, and a figure of merit called
the fill factor has been developed as a single valued ‘quality factor’. Equal
to the product of Vmp and Imp, divided by the product of Voc and Isc, the
higher the value of the fill factor, the ‘squarer’ the IV curve is, and the closer
to ideal the device is. The fill factor in practice, even for photodiodes with
very low Rs and very high Rp, cannot reach the maximum of 100% since
the ideality factor of the diode itself contributes to the rounding of the IV
curve. The departure from ideality arises from various leakage mechanisms
in the p/n junction, some inherent to semiconductors operating above absolute zero, and others due to undesired impurities and imperfections in the
crystalline material.
As mentioned above, concentrator systems have output characteristics
that are very similar to standard PV systems, although the particular cells
used have somewhat higher fill factors than silicon. Recognizable stringing
systems are used, and for the most part, standard inverters are employed.
Direct irradiance varies more than global irradiance, so the diminished
acceptance angle when compared to flat panels causes concentrator outputs
to be more variable. This can be seen in data presented in Fig. 10.4, where
a change in the direct irradiance (DNI) component is plotted against the
© Woodhead Publishing Limited, 2012
Change in DNI across 15 minutes (%, absolute)
Concentrating photovoltaic (CPV) systems and applications
331
200%
:1
180%
s1
160%
e
ng
ha
Ic
N
I, G
140%
120%
tio
Ra
100%
of
DN
80%
60%
40%
20%
0%
0%
50%
100%
150%
200%
Change in GNI across 15 minutes (%, absolute)
10.4 Change in the direct component of global irradiance vs the
change in global irradiance.
equivalent change in global irradiance (GNI). For the vast majority of
changes in the global value, the direct component is affected to a much
greater extent.
As for all photovoltaic systems, concentrators do not intrinsically have
any hold-over capacity or storage. They produce power when the sun is
shining only, and generally their use is backed up by other generation assets,
for example the grid or a local generation plant. Recent intensive research
into large scale electrochemical storage has started to yield very interesting
results however, and it is expected that economically viable, direct electrical
storage will significantly mitigate this limitation within a few years. For an
introduction to this rich field, the reader is encouraged to read Baxter
(2006).
10.2.3
Maintenance
The most common maintenance activity for a concentrator will be cleaning
the environment-facing optical surfaces, for example the mirrors on an
exposed mirror system. This is because they have a limited acceptance
angle, and will be more sensitive to light scattered by dust deposited on
their surfaces than unconcentrated systems. While this is very site dependent, HCPV companies have found that an average of four times per year
is adequate, and it will be somewhat less for LCPV. Exposed mirrors will
be more sensitive than systems where the external optical element is transmissive, for example, a window. The vast majority of CPV systems are
© Woodhead Publishing Limited, 2012
332
Concentrating solar power technology
passively air-cooled, which means that the only water requirement is for
cleaning. In addition, the systems that present a flat, glass face to the atmosphere can be mechanically cleaned using a ‘squeegee’ – a standard window
cleaning device based on a sponge and rubber wiper blade (as opposed to
being deluge-washed), which further reduces the amount of water used
during operation. With a squeegee, also, the purity of the water is not nearly
as stringent as it is for deluge, where deposits from evaporation have to be
regularly removed by other means. It has been estimated by SolFocus, one
manufacturer of HCPV equipment (where the first optical surface is a
window), that their cleaning needs are less than 15 litres of water per
megawatt-hour.
The majority of the balance of maintenance is associated with the tracker.
Mechanical systems need regular maintenance, but if well designed, this is
limited to a biannual lubrication only.
10.2.4
Energy payback and recyclability
As CPV systems are being used to offset dirtier sources of energy, it is
important to understand exactly how clean they are. CPV systems stand
apart from other forms of solar in two distinct ways. First, they are predominantly assemblies of optics and support mechanisms, with very small
amounts of active photovoltaic material. Per unit area, the amount of energy
used to fabricate the cells (whether silicon or multi-junction) dominates
over the manufacture of other materials, such as glass, aluminum and steel.
For CPV, the small area of cell material compared to the area of system will
drive down energy payback time relative to standard PV. Reich-Weiser
et al. (2008) calculated the energy payback time for the SolFocus SF-1100
concentrator panel at 0.7 years, for example, against 2.2–2.7 years for silicon
and 2.2–3.9 years for CSP under equivalent analysis. Second, again because
concentrators are predominantly an assembly of common materials, they
are highly recyclable, and over 95% of a typical system can re-enter the
manufacturing stream at end of life.
10.3
10.3.1
Characteristics of high concentration photovoltaic
(HCPV) and low concentration photovoltaic
(LCPV) devices and their applications
HCPV-specific characteristics
Optical considerations
High concentration ratios can only be reached by point focus systems, so
the acceptance angle will be approximately constant around the optical axis.
© Woodhead Publishing Limited, 2012
Concentrating photovoltaic (CPV) systems and applications
333
This means that all HCPV systems must track the sun in both azimuth and
elevation, and the smaller the acceptance angle is, the more accurate the
tracking has to be. Acceptance angle, then, is a vitally important figure of
merit for an HCPV system. Moreover, workable acceptance angles are
really only possible by invoking the science of NIO, and practically only by
employing multiple-element optical systems (Winston et al., 2005: 1–45).
Single element optics and those employing classical designs, such as found
on some central tower solar thermal plant, will require two to four times
the accuracy required of a well-designed NIO system.
Acceptance angle also plays a role in the energy harvested by an HCPV
system, commonly quoted as the ratio of generated energy to peak power
(kWh/kWp). Under non-ideal conditions for HCPV, circumsolar radiation
– measured from the sun’s limb out to approximately 10° from the solar
axis – can be significant. Thomalla et al. (1983) calculated that under conditions of high cirrus clouds, for example, circumsolar can be several percent
of the sun’s irradiance at 0.5° from the axis. Energy harvest will be lost
under these weather conditions if the circumsolar irradiation falls outside
the acceptance angle. In general, HCPV systems are considered to be sensitive to the direct irradiance or beam radiation only (Lorenzo 2002: 905),
though in practice all have some circumsolar performance. The term used
to describe the beam energy is direct normal insolation (DNI), usually
expressed in kWh/m2-day.
Two-axis tracking
Given that HCPV systems must track the sun in two axes, they do not suffer
from ‘cosine loss’, the decrease in output that comes from energy striking
the cell at large angles from the normal. This effect decreases the energy
harvest from fixed tilt flat panels, with power production approximating a
cosine output with the maximum centered local solar noon. CPV, on the
other hand, has a broad daily output characteristic which is extremely useful
towards the end of the day, as it is usually coincident with some of the
highest daily electricity demand times. To counteract this, there are examples of silicon panels mounted on two axis trackers, but their lower output
has made this a difficult economic proposition. More commonly, the more
efficient silicon panels are effectively mounted on single-axis trackers, with
a fixed latitude tilt.
Where pedestal style trackers are used (Fig. 10.5), their shape and spacing
guarantee that there is no place on the field that is permanently shaded.
This has become an important issue with permitting authorities, many of
whom are concerned with minimal land disturbance, and whether native
vegetation can be re-established after construction. Linear trackers, used in
both the CPV and concentrating solar power (CSP) industries, and
© Woodhead Publishing Limited, 2012
334
Concentrating solar power technology
CPV array and
support framework
Torque tube and
horizontal
(elevation) axis
support
Horizontal
(elevation) axis
actuator
Pedestal
Vertical (azimuth) axis
motion support and
actuator
10.5 Pedestal or azimuth-elevation tracker.
stationary rack-mounted systems will cause permanent shading in some
areas, and disturb the local ecosystem in a significant manner.
A downside to two-axis tracking is that, to minimize early morning and
late evening shading by one system of another, they should be spaced
further apart than stationary installations. Ground cover ratios (the ratio of
the total area of PV to the area of land on which it is installed) of between
17% and 22% are common. While less than the 30% to 50% found in stationary systems, because of high CPV output, this still results in very competitive land use.
Multi-junction cells
HCPV systems can uniquely exploit the recently developed technology of
multi-junction cells. An overview of these cells follows.
As mentioned above, many semiconductor materials are photo-active,
and when used to form photodiodes (or ‘junctions’), can produce useable
power. The bandgap of the semiconductor determines the part of the solar
spectrum the cell is sensitive to, which in turn ultimately dictates the
maximum efficiency.
© Woodhead Publishing Limited, 2012
Concentrating photovoltaic (CPV) systems and applications
335
To break the barrier of single-photodiode efficiencies, research into a
stack of electrically series-connected photodiodes with each photodiode or
junction composed of a different semiconductor material has been underway for several years. The top junction – the one the light sees first – is of
the largest bandgap material, absorbing and converting only the highest
energy photons to power. Photons with a lower energy are transmitted
through the first junction to the next, which is composed of a material with
a lower bandgap. This structure is repeated until the entire solar spectrum
is covered. If the junctions are constructed in a way that their photocurrents
are identical, the total power is simply the sum of the power generated by
each junction. Theoretical work shows a fundamental limit to efficiency of
around 65% (Kurtz and Geisz, 2010: 75) for stacks of this type.
Practical multi-junction cells currently have three junctions, and with
some exceptions, the different designs use very similar materials. The first
junction, constructed epitaxially, is of InGaP and has a bandgap of 1.8–1.9 eV;
the middle junction is also epitaxially constructed and is of GaAs with a
bandgap of 1.3–1.4 eV. The bottom junction is usually diffused into Germanium and has a bandgap of 0.7 eV. Efficiencies have climbed dramatically
over the past few years, from 25% in 1990 to 42% in 2010 (Kurtz and Geisz,
2010: 74), and recent work on cells having greater than three junctions is
beginning to yield results over 40%.
In addition, their efficiency is relatively independent of their operating
temperature. Unlike silicon cells, with coefficients of around −0.4%/C
(Green, 1998), modern multi-junction cells can be in the range of −0.07%/C
(Spectrolab 2010). Because of this, to a very large extent, the cooling system
in a concentrator is designed to maintain the cell within its specified operating temperature range only, and not to maximize output. CPV fields show
little output degradation with increasing ambient temperatures, and they
are economically very attractive for hot, dry desert conditions, where
daytime temperatures can reach over 50°C.
Nothing comes for free, however, and several intrinsic issues must be
addressed with any of these complex cells. First, if the different junctions
are stacked on top of one another, a parasitic diode will be formed between
them. These parasitic diodes will seriously impair the performance of the
overall structure, and must be eliminated. This is done by inserting tunnel
diodes between the main junctions, isolating the individual photodiodes.
Second, when correctly biased at their maximum power point, each junction operates as a current source. Because of this and their series electrical
connection, they must be sized to generate identical photocurrents, as the
resulting current from the stack will be equal to that of the junction with
the least current generation. This makes them more spectrally sensitive than
single-junction cells, and since the sun’s spectrum is not constant during the
day, it is inevitable that most of the time one or more of the junctions will
© Woodhead Publishing Limited, 2012
336
Concentrating solar power technology
underperform and dictate the device’s photocurrent. Because of this, welldesigned CPV systems will exhibit a daily efficiency variation of a few
percent, and significant design attention must be paid to the concentrator
optical transfer function, since it modifies the incoming energy spectrum
(McDonald and Barnes, 2008). It is worth noting here that recent research
into the bandgap-widening effect of nano-structures, for example quantum
dots or quantum wells, has met with success, and will serve to lower the
spectral sensitivity of the cell, and increase overall energy harvest.
The above two concerns, along with other practical crystallography issues
result in a very complex device with long technical development times, and
ultimately a high intrinsic product cost. The output from a multi-junction
cell is large, and can be exploited economically effectively only by HCPV
systems.
It is worthwhile underscoring one final point, briefly visited above.
Despite years of development and reliable operation in space, multijunction cells are still technically young, with room for performance
improvement. Currently commercially available cells have efficiencies of
approximately 38.5% at 25°C, and some companies are demonstrating cells
at around 42%. Notwithstanding the spectral issues mentioned above, enormous gains can be made in performance, and several reputable firms have
three- to five-year product roadmaps that border on 50% efficiency. Indeed,
the history of the multi-junction cell is one of steady, concerted performance gain year on year. Efficiency increases are a very strong lever in the
cost of HCPV systems. Not only do the concentrators themselves become
less expensive for a given power output, but with fewer systems to install
per megawatt, balance of plant and operating and maintenance (O&M)
costs will also decrease.
10.3.2
LCPV-specific characteristics
As mentioned above, LCPV systems occupy concentration ratios of between
1.25 and 40. Because of their low concentration, they can use designs that
have lower achievable maximum concentration ratios, like parabolic troughs
(Swanson, 2002: 479–482), linear Fresnel reflectors, the compound parabolic
concentrator (CPC), (Winston et al., 2005: 50–89) or the V-trough (Sangani
and Solanki, 2007). These two-dimensional concentrators can be thought of
as concentrating in one axis only, the other axis operating at one sun. In
addition, their acceptance angle in the concentration axis is large compared
to HCPV systems.
If the concentration is very low, for example around 2–2.5 suns, the
acceptance angle in the constrained direction will be above 23.75°, which
is the tropical latitude plus the angular radius of the sun. If the concentrator is positioned carefully, the sun’s arc during the entire year will fall
© Woodhead Publishing Limited, 2012
Concentrating photovoltaic (CPV) systems and applications
337
within the acceptance angle, and the system will not need to track. For a
theoretical treatment, see Luque (1989: 305). For cost reasons, these systems
are usually constructed as V-troughs or possibly as a compound parabolic
trough.
As concentration increases from 2.5 to, say, 10 suns, concentrators can be
tracked by making seasonal adjustment only. Twice per year they are moved
so that for the following six months, the sun’s arc will fall within their
acceptance angle of down to 12°. Continuing from 10 suns, the tracking
requirements become increasingly more stringent, but still needing movement in one axis only.
Another consequence of the high acceptance angle is the ability of the
LCPV system to operate with circumsolar and, in the very low concentration systems, most of the diffuse light. As a result, an LCPV system will have
a smoother output response than an HCPV system under intermittently
cloudy conditions.
For all of the above, there is a significant challenge to the design, fabrication and operation of an LCPV system: cost. Because of the low concentration levels used, LCPV systems cannot overcome the cost hurdles presented
by high performance multi-junction cells, and so use silicon cells. The
lowered output in turn demands that very inexpensive optics, thermal management and tracking systems be employed, since there is not as much
power over which the support structures and mechanisms can be amortized.
In principle, any low cost photovoltaic technology, for example thin film,
could be used, but achievable optics and tracker costs have still mandated
using high efficiency mono-crystalline silicon. In addition, because of the
lower output, there are constraints on the amount of effort that can be
expended on system maintenance, with the result that the very inexpensive
systems also have to be quite reliable.
10.3.3
Medium concentration photovoltaic devices (MCPV)
From the interest shown at both ends of the concentration scale, there is
ample evidence that viable products can be built and operated. There is no
evidence yet of domination of HCPV over LCPV or vice versa, but more
of differentiation into market segments. Interestingly though, there appears
to be little work being done on equipment with concentration ratios between
50 and 200.
At the low end, parabolic troughs can operate above 40 suns, but they
have a theoretical maximum concentration predicted by NIO to be around
100 suns and as they reach this limit, tolerances and accuracy requirements
render them expensive, but with an output too low for effective amortization. In the solar thermal space, the mechanical and hydraulic advantages
of troughs have caused designs to be refined around 80 suns, though there
© Woodhead Publishing Limited, 2012
338
Concentrating solar power technology
is still a practical and cost limit significantly below what is theoretically
possible.
At the high end of the spectrum, dual-axis tracking and multi-junction
cells are the norm. As concentration decreases below, say, 250 suns, systems
still need the trackers for accuracy, but the decreased power density increases
the cost per unit power. At some point the high cost of the cell cannot be
justified and the switch to silicon must be made. But with the lowered
output, the two-axis tracker cannot be justified. One concentrator company
operated a point focus system at around 150 suns with silicon for several
years, but economics caused them to increase the concentration and migrate
to a triple-junction cell.
In summary, the occupancy of both ends of the concentration landscape
is due mainly to cost optimization. HCPV systems have sufficient output
over which to amortize their relatively complex designs, while LCPV
systems rely on non-stringent requirements for their optics and trackers,
leading to low cost.
Now with a basic understanding of the characteristics of concentrator
systems, their place in the market can be assessed.
10.3.4
Application to the market
The viable solar market spans an enormous range of geographies and
weather systems, significantly complicated by access to transmission assets,
finance, construction resources and power users or offtakers. Government
at all levels and semi-government institutions, such as utilities add a further
and important complication. In addition, while the residential rooftop
market dominated the growth of the solar sector for many years, as the
reliability of photovoltaics has been proved and costs shown to be increasingly attractive, interest in large installations has become common. Usually
ground-mounted and increasingly in hot, dry areas of the world (e.g., around
the Mediterranean), solar power stations of up through hundreds of megawatts are now under development, and will be in operation well before the
end of the decade. The result is a rich set of competitive ecosystems that
favor one or another of the many technologies available today. The question
is: Where are concentrators appropriate?
The very low concentration devices – those that don’t track – will find
use alongside standard silicon panels. Their economic arguments usually
center around panels with outputs similar to standard panels, but at lower
cost because they use less silicon. Decreased acceptance angle when compared to standard panels makes them more sensitive to light scatter, and
hence they will have to be cleaned more often, and they will be of limited
use in very hazy conditions. This will apply pressure for larger installations,
where the operation can be amortized over a larger number of panels, for
© Woodhead Publishing Limited, 2012
Concentrating photovoltaic (CPV) systems and applications
339
example on commercial rooftops. In addition, because they don’t need to
track, low concentration devices are probably the only viable concentrator
for rooftop applications. A small number of companies have been developing higher concentration systems for rooftop application, but the economics
have been shown to be risky, especially in the urban environment with
significant levels of atmospheric scattering. The relatively low output of the
stationary systems means they will be less competitive against higher concentration devices, so they will probably not find large use in field-mounted
situations.
Tracking LCPV systems – those from 10 to 40 suns – have the advantage
over HCPV in that they are less sensitive to manufacturing tolerances, but
are disadvantaged in output density. While having higher output than stationary concentrators, their mechanical complexity and maintenance
requirements will probably prohibit them from rooftops. Their output levels
combined with wide acceptance angles, though, will make them ideal for
distributed generation in the urban environment. Maintenance requirements and the complexity associated with installing mechanical systems will
apply commercial pressure to install tracking LCPV systems in fields of at
least tens to hundreds of kilowatts, so that O&M tasks can be carried out
cost effectively. At the high end of LCPV, care must be taken with location,
as the acceptance angle, combined with the pressure for very low manufacturing cost, is sufficiently small as to disadvantage them in areas of high
diffuse radiation. Like HCPV, the top end LCPV systems will be limited to
areas of high DNI.
HCPV systems have the highest power density of any photovoltaic technology. They also require the most complex, expensive tracking systems, and
have the lowest acceptance angle. HCPV equipment will be economically
advantageous when aggregated into power stations of hundreds of kilowatts through to many megawatts, and will be operated in areas of high
DNI – the ‘sun belt’ regions. Specifically, their high power output under
elevated ambient temperatures coupled with very small water use (and, in
many cases, the ability to use untreated water), means that HCPV is ideally
suited to hot, dry climates. This is a market niche for CPV which is, however,
very large, and growing fast. Market research shows this to be approximately 50% of the ground-mount solar marketplace, or 20% of the total
available market.
10.4
Design of concentrating photovoltaic (CPV)
systems
The preceding discussion on the characteristics of the technology and its
entry points to the market now sets the stage for an introduction to the
design elements and challenges for a concentrator. The basic issues at the
© Woodhead Publishing Limited, 2012
340
Concentrating solar power technology
heart of any CPV design are introduced in this section, and a small set of
example products are described.
10.4.1 Levelized cost of energy
Ultimately, a CPV plant is a financial instrument, and investors will make
technology and project partner choices based on a return analysis. In addition, since a solar plant provides an annuity throughout its lifetime, any
investor will require an accurate assessment of its ongoing financial performance against expectations. An effective metric that provides both of these
needs is the levelized cost of energy (LCOE). LCOE has been introduced
in Chapter 2; it is an important metric for CPV system design as it is with
any CSP technology (Short et al., 1995). Fundamentally, it is a calculation
of the lifetime cost of energy for any generating plant, brought back to net
present value and includes all fixed and marginal costs, degradation in
output, projected repair and replacement costs, and inflation and discount
rate estimations. Expressed as ¢/kWh, this type of calculation can uniquely
be used to compare technologies, for example CPV against PV, hydroelectric or fossil fuel generation. Short et al. (1995) define LCOE as ‘that
cost that, if assigned to every unit of energy produced (or saved) by the
system over the analysis period, will equal the Total Lifecycle Cost (TLCC)
when discounted back to the base year.’
The formal definition for LCOE, if tax issues are not considered, can be
expressed as:
LCOE =
NPVLCC
∑
N
1
(Ei /(1 + DR)i )
[10.3]
where Ei is the energy output in year i, N is the amortization period, DR is
the discount rate, and NPVLCC is the net present value of lifecycle costs,
which is given by:
NPVLCC = ∑ i = 1
N
Ci
(1 + DR)i
[10.4]
where Ci is a cost during period i.
For a CPV system, the cost cashflows can be categorized as (Nishikawa
and Horne, 2008):
•
•
•
the cost of installing the CPV/concentrator system which will essentially
be a single investment in year zero
the cost of installing balance of systems (BOS), i.e. inverters, civil works,
etc., will also be a single large investment in year zero
the annual operation and maintenance cost.
© Woodhead Publishing Limited, 2012
Concentrating photovoltaic (CPV) systems and applications
341
Bringing back all costs and generation to net present value allows the comparison of different technologies with different financial structures and
lifetimes to be made at the time of a potential investment.
The quantification of energy and not operating power is central to the
theme of LCOE. While it is common to use installed capacity-centric costs
($/W or $/kWe) when discussing solar plant of all types, it has two very
important drawbacks. First, the rating schemes for the various technologies
are not the same. As explained by Melia et al. (2010), because of the way
the rating conditions are defined, a 300 W HCPV panel will actually produce
300 W at 20°C and under 850 w/m2 direct irradiance (the rating irradiance).
The rating conditions are different for silicon panels, however, with the
result that an identically labeled poly-silicon panel will produce 264 W
under its rated condition of 1,000 w/m2 global irradiance. These differences
are due to test practicalities, but if used in the design phase, will lead to
power plant of very different sizes being built, with different up-front
investments and different energy harvests.
Second, the specific energy, a metric used to indicate the ability of a
technology to harvest energy, and defined as the ratio of kWhannual to kWrated,
varies considerably by technology. There are many reasons for this, including whether a technology is mounted on a tracking system, the size of the
temperature dependence on output, geographic spacing of systems, the
acceptance angle and their susceptibility to soiling. As an example, a CPV
system with a very narrow acceptance angle might produce as much power
when directed on the sun as one with a wider one, but over the course of
the year, the larger acceptance angle system will admit more circumsolar
radiation, will be less susceptible to tracking errors, alignment issues and
foundation movement. The product with the larger acceptance angle will
harvest more energy, returning a greater annuity to the investor. So, using
a power-based metric tells very little of the investment story, and eliminates
the ability to choose between technologies or project equipment
suppliers.
LCOE is gaining ground as a standard metric, largely due to the US
Department of Energy and NREL investing significant effort in publicizing
its value and producing tools for its use. The reader is advised to investigate
the Solar Advisor Model (SAM) from NREL, and explore its supporting
documentation (Gilman et al., 2008).
10.4.2
General system design goals
Generally, the overall design goal of a CPV system is to produce a product
that minimizes the LCOE in the chosen geography. Within that simple
statement, however, lie many challenges and trade-offs. The first is that of
scale.
© Woodhead Publishing Limited, 2012
342
Concentrating solar power technology
To a first order approximation, regardless of efficiency, all PV technologies need to be fabricated and deployed in very high volume. Consider a
30 MWPdc-rated CPV power plant with equipment operating at 27% efficiency, and a concentration ratio of 500 suns. At that efficiency, the input
power would have to be 1.1 × 108 W to satisfy the output requirements. At
the 850 w/m2 irradiance rating, 1.3 × 105 m2 of concentrator area would be
required which, at 500 suns, means 2.6 × 102 m2 of cell. A common size for
a concentrator cell is 7.5 mm on a side, for approximately 0.56 cm2, so 4.7
million of these cells would be installed in the plant. Each of these cells will
require electrical connections, cooling, alignment, a share of the optical
system and eventually to be integrated on a tracker of workable dimensions.
Given the sheer numbers associated with this moderately sized plant, it
should be clear that very high volume manufacturing, in the scale of
the automotive industry, for example, is needed for successful CPV
deployment.
Luckily, most photovoltaic concentrator designs, while seemingly
complex, lend themselves exceptionally well to standard high volume
manufacturing techniques as used, for example, in the automotive or electronics industries, with their attendant low costs of production. They are
essentially an assembly operation, as opposed to thin film, for example,
where thousands of square meters of complex vacuum deposited material
must be produced.
Further, despite the complexity of CPV systems, they must be able to
operate within accepted reliability norms for solar, usually guaranteeing
80% output after 25 years. A very good understanding of degradation
mechanisms at work in any concentrator design is of paramount importance, as is the ability to accurately measure them. While this might seem
to be a difficult task for a new technology, two things must be considered.
First, because concentrators are assemblies, their components and assembly
methods can be designed and tested separately for reliability. Second, if
designed well, many of the subsystems use techniques, materials and assembly methods adopted from other industries with years of experience. Test
protocols, design advice and degradation data are all available to guarantee
a reliable system. An aluminum drawn backpan for a concentrator, for
example, can use techniques from the automotive industry where simulations, materials choices and test techniques can all guarantee 25-year
longevity.
Finally, performance is important, and the paramount concept is, as mentioned above, minimizing the LCOE of the product. This metric is informed
by all other performance parameters and in turn will inform the financial
performance of the field.
There are many degrees of freedom with concentrator design, and as a
result, many different styles of product have emerged. So many, in fact, that
© Woodhead Publishing Limited, 2012
Concentrating photovoltaic (CPV) systems and applications
343
categorization has been virtually impossible beyond those of HCPV and
LCPV. Adding to the confusion is the fact that there are still many unexplored design pathways. Observing some trends, however, one can make
some necessarily imprecise generalizations. The rest of this chapter will
describe these general categories. A small set of examples will be used for
illustration.
System granularity
System granularity refers to the overall architecture, and whether the concentrator is composed of an array of small concentrators (usually each
irradiating a single cell), or a single, large optic, powering an array of cells.
There are instances of both approaches, covered in the examples below. An
array of small concentrators is by far the most common approach, as it lends
itself better to modern manufacturing techniques.
Optical method
The optical method may be refractive, reflective, or catadioptric (which
employs both refractive and reflective techniques). The former two are the
most common, with refractive Fresnel systems being favored by many.
Cassegranean reflectors are being pursued by a smaller number of companies, as are single reflective optics. Catadioptric non-imaging optical elements are sometimes used as the last stage of a complex concentrator. One
example of a complete concentrator of this type was designed and prototyped at Universidad Polytechnic Madrid (Swanson, 2002: 492–494) though
it has not been commercialized. The optical design will drive all other
aspects of a concentrator design, from thermal management through to the
manufacturing processes that come available.
Tracking type
There are many mechanical styles of tracking systems. Degree of concentration will largely determine whether single- or double-axis tracking is chosen.
Beyond this, there are not a strong set of criteria. Each appears to have
advantages and disadvantages, none of which appear to be overwhelming.
Tracker types generally fall into three broad categories:
1. Azimuth/elevation tracking systems (see Fig. 10.5 above) sit on top of
a pedestal or pole, with one axis rotating vertically (azimuth) and the
other horizontally (elevation). Polar tracking systems are a subset of
this, where one of the axes, usually the azimuth, is inclined at the local
latitude angle.
© Woodhead Publishing Limited, 2012
344
Concentrating solar power technology
Horizontal axis (roll)
support and actuator
Vertical axis (tilt)
actuator
CPV array and support
framework
Vertical axis(tilt) ganged
movement mechanism
Horizontal axis (roll)
support
Pedestal
10.6 Tilt-roll tracker.
2. Tilt/roll systems have one horizontal axis and a second, attached to the
first, at a right angle to it. An example of a tilt/roll tracker is shown in
Fig. 10.6.
3. Finally, carousel trackers are a special form of the azimuth/elevation
tracker, but are very compact, and possibly suitable for rooftop mounting. They are mechanically complex, however, and present a manufacturing and tolerancing challenge.
There are multiple means to control or steer tracking systems. First,
movement is commonly driven by electric motors, though there are several
examples of hydraulically actuated systems on the market. Control system
algorithms vary from using external sensors, to sampling the output of the
concentrator itself, to dependence on an ephemeris equation. While there
remains a great deal of variation on the market, a combination of ephemeris
plus monitoring the concentrator output appears to be the most practical
method.
Trackers are a complex topic, especially when the issues of wind, local
building codes, structural vibrations, logistics, deployment, and lifetime are
considered. Details are, unfortunately, beyond the scope of this single
chapter on CPV.
Environmental control methodology
CPV systems need to include thermal management of the cell. Despite the
high efficiencies of multi-junction cells, currently still over 50% of the
© Woodhead Publishing Limited, 2012
Concentrating photovoltaic (CPV) systems and applications
345
incident energy is converted to heat, which must be removed in order to
maintain reliability. In general the trend is towards passive, dry cooling,
where the thermal energy is conducted away from the cell, then radiated
and convected to the atmosphere. Active air cooling schemes have been
tried on some systems, but generally the need for reliability and low maintenance costs mandate against them. Concentrators with larger optical
systems and those with very large concentration ratios generally must use
active cooling, which can include heat pipes and fluid reticulation.
Additionally, there is a need to protect the optical systems from the elements. Fresnel lenses with their fragile teeth and generally deep valleys
would lose efficiency rapidly if employed in the outside environment. They
are almost universally enclosed within a housing, where the front window
is flat, and the lens elements are moulded or embossed onto the inner,
protected surface. Reflective systems have more leeway, and several designs
have external mirrors, in the same manner as solar thermal technologies.
Several reflective technology companies still choose to house the mirrors,
however, further protecting them.
In general the environment inside the housing is maintained through the
use of passive air filtration, though active drying techniques have been seen.
Cell management
The immediate area around the cell itself is a technical and financial challenge. Within a few millimeters of the cell, thermal management must be
carried out, robust electrical connections made, a bypass diode mounted,
mechanical alignment carried out, and possibly very high incident solar flux
managed, regardless of the alignment of the concentrator. There have generally been two approaches to these multiple challenges. The first has been
to simply mount the cell or an assembly consisting of the cell and bypass
diode on the concentrator superstructure, and complete the construction
around it. In the second approach, a ‘receiver’, which consists of the cell or
cell array and related thermal, electrical, and optical components, is assembled separately and mounted as a subassembly to the concentrator in an
independent manufacturing step.
Both approaches have their advantages, generally in the manufacturing
arena, and there are examples of both in the market.
10.5
Examples of concentrating photovoltaic (CPV)
systems
Following are examples of some of the CPV systems that have been built
in significant volume or are under development.
© Woodhead Publishing Limited, 2012
346
Concentrating solar power technology
Target:
1. Cell array
2. Thermal management
Parabolic
mirror
Pedestal
for Az/El
tracker
Support spider:
1. aligns target
2. access for cooling fluid
3. access for power output
10.7 Point focus, imaging paraboloidal concentrator.
10.5.1
HCPV single dish reflective
Many early HCPV concentrators were built to a design common to optical
and radio telescopes: the point focus, imaging paraboloid (Fig. 10.7). Solar
Systems of Australia has commercialized one version, a large single reflecting optic of approximately 15 m in diameter, targeting an array of watercooled multi-junction cells (Fig. 10.8). Several significant installations have
been operating for some years in the center of Australia, powering remote
grids and demonstrating the efficacy of the approach.
These systems are optically highly efficient, because they use only a single
reflective element to achieve the necessary concentration. In addition, the
optics are external, and not enclosed within a housing, so there is no transmission loss through a front window. Being reflective, they do not suffer
from chromatic aberration. This architecture, though, while delivering high
peak power, suffers from a limited acceptance angle. Because it is a singleelement imaging system, as shown by Welford and Winston (1989: 31–51),
the attainable acceptance angle for a given concentration will be much less
than that of a non-imaging design, practically between 25% and 50% of
what is possible. The system, then, needs to very accurately track the sun,
and will be susceptible to atmospheric scatter.
Energy conversion by an array of cells mandates the use of liquid cooling,
but also allows it, since relatively expensive pumps, filters, valves, radiators
© Woodhead Publishing Limited, 2012
Concentrating photovoltaic (CPV) systems and applications
347
10.8 Example of a point focus imaging paraboloidal concentrator from
Solar Systems, Australia.
and piping can be amortized over a large power output. This is in contrast
to the array-based systems described later, which, with an optical element
per cell, are limited to mainly passive cooling. The parasitic energy required
by the active cooling system, however, can rob several percent of the harvested energy from the system. While less than the 10% estimate for most
thermal systems, this loss is significant. There are further compromises to
consider also.
First, the interaction of the multi-cell target, and the inherently nonhomogeneous illumination received from the reflector will decrease output
due to mismatched cell photocurrents between the different cells in each
series connected string. This effect can be minimized by the use of a blocking diode per cell, but cannot be eliminated. This spatial variation in power
delivered to the cells will vary over time as the system tracks the sun, so an
optimum electrical connection architecture for the cells is difficult to realize.
Second, the cells produced by all current manufacturers include a busbar
on the light-facing side of the cell, for the negative terminal. There are no
back-side contacts, as in some high-performance silicon cells. The area presented by the busbar and associated electrical connection does not itself
© Woodhead Publishing Limited, 2012
348
Concentrating solar power technology
generate electricity, and represents a harvest loss that can be higher than 5%.
This issue can theoretically be minimized by the use of reflectors on top of
the cell busbars, but to the author’s knowledge, this has not been tried.
Finally, the exposed mirrors will lose more power for a given amount of
soiling than will any system where the external interface is a transmissive
element, for example a Fresnel lens or a flat window (Vivar et al., 2008).
This is because not only will a particle of dust shade incident light, it will
also scatter light reflected from a nearby area of the mirror.
Field examples use segmented mirrors to improve installation viability,
and the mirrors are fabricated from glass, metalized on the second surface.
This helps with the need for high precision optical alignment processes in
the field, but by no means eliminates this somewhat slow and skilled task.
Mirrors constructed in this manner for the solar thermal industry can withstand years of mechanical cleaning, but especially with the new non-metallic
protective layers mandated by environmental laws, still have questionable
longevity when used outdoors (Kennedy et al., 2007).
The remainder of the system consists of a space-frame mirror mount
attached to an azimuth/elevation tracker.
So, while optically efficient, energy harvest in a single dish reflective
HCPV system presents a number of issues, its acceptance angle is quite low,
and it will be more sensitive to contamination than most other types.
10.5.2
HCPV Fresnel lens array
In this approach, a Fresnel lens (either refractive or catadioptric) is mounted
on a housing to entrain light to a photovoltaic cell or cells. Figure 10.9 shows
a classic refractive Fresnel lens design. The reader is directed to Leutz and
Suzuki (2001) for an excellent treatment of the complete genre of Fresnel
lenses. The design is straightforward, though because the optical path is not
folded (a possibility only if using reflectors), it yields a deep system with
somewhat stringent mechanical tolerance requirements.
The vast majority of Fresnel systems are constructed as assemblies of
small concentrators within an enclosure, with one cell per optical system.
Usually the front of the enclosure is the lens or a series of lenses and
support members. While the lenses themselves can be fabricated in large
geometries, their size is chosen in consideration of the target concentration,
the cell size and its thermal management requirements, and the maximum
reasonable depth of the enclosure. Using one cell per optical assembly
minimizes problems encountered from inhomogeneous illumination. The
cells themselves can withstand a surprising degree of non-homogeneity with
low efficiency loss (Katz et al., 2006), and it is easier to balance the radiation
on each cell in a string through the use of a dedicated optical system than
by illuminating an array of cells from a single optic.
© Woodhead Publishing Limited, 2012
Concentrating photovoltaic (CPV) systems and applications
349
Fresnel lens
Target
(single cell)
10.9 Cross section, refractive Fresnel lens concentrator.
The main benefit of a concentrator of this type is its high optical efficiency
through the use of a single optical element, or two in the case where a nonimaging final stage is used. However, while a simple system, the Fresnel lens
does exhibit losses that are hard to overcome. First, the teeth of the Fresnel
system have to be made very accurately in order to achieve high concentration. Rounding at the teeth tips and filling in of the intervening valleys due
to manufacturing imprecision can cause light scattering, and mechanical
requirements for efficient mold operation can cause obstructions in the light
path. In addition, an anti-reflective coating can be placed only on the flat
part of the Fresnel and not the toothed surface, which contributes a further
energy loss through unwanted reflection.
These relatively simple, high-efficiency optics do not yield the best acceptance angle, so these concentrators, while possessing high peak power, need
to track the sun quite accurately to minimize output variations. The acceptance angle limitation also means they are more susceptible to large particulate aerosols, for example those found in cirrus clouds.
In addition, chromatic aberration, in which the refraction of the light
through the lens is frequency dependent can lead to losses if not properly
© Woodhead Publishing Limited, 2012
350
Concentrating solar power technology
managed. The effect of chromatic aberration is to alter the spectral balance
on the cell, with a resulting energy conversion loss.
Most of the above problems can be minimized through careful design.
Swanson (2002: 488) shows a domed lens that eliminates the tip losses, for
example, and Leutz and Suzuki (2001: chapter 6) illustrate many methods
in great detail. There are several Fresnel systems that minimize these problems, and lens efficiencies approaching 90% are possible (Leutz and Suzuki,
2001: 118).
Fabrication of the lens has split into two camps: poly(methyl methacrylate) (PMMA, or acrylic) and silicone-on-glass (SOG). PMMA has the
longest history and is the easiest to fabricate, usually by embossing the lens
elements onto a flat sheet of PMMA. Good mechanical tolerances can be
achieved, and the process is fast, and scalable to high volume production.
A PMMA lens, though, suffers from surface and internal degradation
(Rainhart and Schimmel, 1974), and is known to be susceptible to damage
from mechanical cleaning (e.g., with a squeegee) and stain-causing airborne
pollutants.
SOG lenses are fabricated by casting the Fresnel lens elements onto one
face of a sheet of glass using a clear silicone gel. Silicone bonds well to glass,
and the mechanical structures formed are at least as accurate as with
PMMA. In addition, the lens has a durable outside face of glass and does
not suffer the degradation problems of PMMA. While questions still exist
about the robustness, scalability and ultimate cost of lenses built this way,
they are becoming more common in Fresnel concentrators.
The balance of the concentrator is made by forming a backpan, which
contains the cells or receivers, alignment mechanisms for the lens array, any
environmental control (e.g., filters or air dryers), and the means for attachment to a tracker. While it is usually fabricated from metal, at least one
backpan has been prototyped from injection-molded plastic, and a system
with an all-glass backpan is undergoing commercialization.
A significant advantage of the optical array architecture is that the majority of the high precision operations are carried out in the factory and not
in the field. The receiver’s size and components lend themselves to assembly
on standard electronic assembly lines (including the use of high-speed pickand-place equipment). The low number of part types intrinsic in an arraybased design also allow for highly automated assembly, and automotive
style manufacturing lines using six-axis robots are becoming common. The
result is that completely operational, tested modules are shipped to the
field, requiring a minimum amount of in-field alignment. Installation is possible by supervising locally sourced labor with a minimum of training.
Depending on the ultimate size of each module, their installation on the
tracking system can be done with locally sourced equipment.
© Woodhead Publishing Limited, 2012
Concentrating photovoltaic (CPV) systems and applications
351
10.10 Example of a refractive Fresnel lens concentrator from
Concentrix GmbH, Germany.
Fresnel concentrators are the most common of all, with the most mature
companies being able to show several years of field experience. Concentrix
GmbH, a spin-out from the Fraunhofer laboratories is one such leader, with
their advanced silicon-on-glass FLATCON© concentrator. Figure 10.10
shows one module of their system. Several of these modules are integrated
on the support frames of their proprietary dual-axis tracker to produce a
system of approximately 6 kW rated power. Concentrix has installed a
significant number of systems globally, and operates a 25 MW/year manufacturing plant.
10.5.3
HCPV complex reflective
Cassegrain optical geometry can be used to produce a concentrator system
with very compact geometry, with wide acceptance angle and good efficiency (Fig. 10.11). In this design, a converging primary mirror entrains light
on a diverging secondary, which in turn focuses on the entrance to an NIO
final stage concentrator. Having the optical path ‘folded’ in this manner
produces a system that can be more than three times more compact than a
Fresnel system for a given input optic size. In general, mirrors are more
efficient than refractors, and make possible the introduction of a larger
© Woodhead Publishing Limited, 2012
352
Concentrating solar power technology
Secondary mirror
Primary mirror
Non-imaging
tertiary
Target
(single cell)
10.11 Cross section, reflective cassegrainean concentrator coupled to
a non-imaging tertiary.
number of optical surfaces between sun and cell without large power loss.
This increases flexibility when making efficiency, energy harvest, and manufacturability tradeoffs, over that of simpler systems.
Like Fresnel systems, this approach lends itself to being constructed as
assemblies of small concentrators with one cell per optical system and with
the use of prefabricated receivers. The optical system’s size is also chosen
for reasons similar to the Fresnel: concentration, cell size and thermal management technique. The folded optical path greatly reduces the bulkiness
of the assembled system, which benefits assembly, logistics, and fielding.
With careful design, excellent efficiencies can be achieved, approaching
what is possible with simpler systems. The larger number of optical surfaces
directly trades against the throughput efficiency limits of the Fresnel lens
and in principle the difference can be small. In practice, however, since very
high quality reflective surfaces are expensive, Fresnels are usually somewhat more efficient than cassegrain reflectors, at least when considering
only the transport of light through the optical system. Because of the use
of reflection as opposed to refraction, however, the cassegrainean system
delivers its efficiency with barely a trace of chromatic aberration and, especially at high concentration ratios (e.g., 1,200 suns), this can greatly mitigate
the throughput loss as the spectral/spatial characteristic of the output light
is better matched to the cell.
In addition, the complex optics allow acceptance angles reasonably close
to the practical limit, and within its acceptance angle the output can be
© Woodhead Publishing Limited, 2012
Concentrating photovoltaic (CPV) systems and applications
353
made more uniform than that of more simple optical systems, which exhibit
a more sinusoidal response over angle. Gordon (2010), for example, illustrates two reflective surfaces being tailored to produce an aplanatic system,
that greatly minimizes coma, an optical aberration that deteriorates acceptance angle. As a result, a cassegrainean HCPV system will be less susceptible to scattering by suspended aerosols, and will be less sensitive to
manufacturing tolerances and moderate tracking errors.
It is also worth noting that with the array approach, the reflectors are
generally placed behind the window of a housing, rendering a flat, easy to
clean surface and further decreasing the susceptibility of the concentrator
to soiling.
The main drawback is the complexity of the system, which mandates a
refined approach to volume manufacturing, including a well thought out
automation strategy. While many prototype cassegrainean concentrators
have been built over the years, a marriage between design and automated
processes has only recently been successful.
Reflectors can be constructed of many different materials and using many
techniques, since light need not necessarily be transmitted through the
optical element. Importantly, in order to transmit the entire spectrum the
multi-junction cell is sensitive to, there are a limited number of practical
options for the reflecting material, which usually employs silver, or spectrally enhanced silver or aluminum. There are other significant tradeoffs in
the combinations of materials and techniques, and their importance depends
somewhat on the overall design. There is not, currently, a single obvious
methodology. Table 10.2 illustrates the advantages and disadvantages of
some of the more common materials and forming techniques.
As in practical Fresnel systems, the design is rounded out with a backpan.
Because of the compactness afforded by a folded optical path, it can be
fabricated from a simple stamping operation, yielding an inexpensive and
robust subassembly. All of the manufacturing advantages associated with
array-based architectures described above apply, and multi-megawatt/yr
manufacturing plants have been shown to be inexpensive and fast to start
up. Operating and tested concentrator panels are delivered to the field site,
and the large acceptance angle minimizes requirements for planarizing the
panels on the array frame. This speeds the overall installation process.
The SF-1100 from SolFocus Inc. is one example of a commercially ready
system. Based on slumped glass mirrors, the optics are protected inside a
glass and drawn metal housing. Twenty-eight of these modules are aggregated on a proprietary dual-axis tracker, to yield 9.2 kW under rated conditions (Fig. 10.12). The company has over three years of operation on its
equipment and runs a 30 MW/year, highly automated production line. SolFocus has installed globally, and has shown that locally sourced labor and
equipment can efficiently build a solar power plant based on the SF-1100.
© Woodhead Publishing Limited, 2012
© Woodhead Publishing Limited, 2012
Injection mold
Injection mold
Slumping
Drawing
Polyolefin or
polycarbonate
Glass
Aluminum
Vacuum forming
PMMA
PMMA
Forming technique
Base material
Electro-deposition,
with or without a
base layer of
polymer or glass frit
Rear surface, liquid
deposition
Front surface, vacuum
deposition
Front surface, vacuum
deposition
Front surface, vacuum
deposition
Metalizing technique
Table 10.2 Comparison of reflector fabrication techniques
Lightweight, robust
Hydrophobic
Least expensive
Good Ag adhesion
Hydrophobic
Lightweight
Large array fabrication
Lightweight
Large array fabrication
Lightweight
Advantages
Good surface finish is very
difficult to achieve.
Difficult to fabricate an array of
mirrors.
Most expensive option.
Ag adhesion issues.
Hydrophilic material.
Ag adhesion issues.
CTE mismatch with most
protective coatings.
Needs very good mold quality.
Hydrophilic material.
Ag adhesion issues.
CTE mismatch with most
protective coatings.
Disadvantages
Concentrating photovoltaic (CPV) systems and applications
355
10.12 Example of a reflective cassegrainean concentrator from
SolFocus Inc., USA.
10.5.4
LCPV reflective
There are a modest number of LCPV systems under development, including
the V-trough non-tracking concentrator. This design can be thought of as an
inexpensive silicon panel, and with few restrictions, can be used interchangeably. Because of the similarity to standard panels, the V-trough will not be
considered here. Instead, we will consider a tracking example, a system
similar to that of Benitez et al. (1997) called the dielectric single-mirror two
stage (DSMTS) concentrator, and which is under commercialization.
This device operates at concentration sufficiently low that a point focus
optical system is not needed, and a less expensive linear design can be
employed. In addition, single-axis tracking is sufficient. It consists of two
trough-shaped reflecting elements, arranged to direct light onto each other’s
strip receivers’ mounted reflector edges, as shown in Fig. 10.13. Because of
the low concentration, the optical target area is too large for the use of
multi-junction cells to be economically viable, and high performance silicon
cells are used instead. This is a very low cost approach that allows the reflector to be the alignment and mounting device for the receivers.
With only a single optical element, the optical efficiency will be quite
high, and being a trough, the irradiation on the cells will be uniform in the
lengthwise direction. Because of the low concentration (approximately 40
suns), the acceptance angle will be reasonably large, and the system can be
tracked on an inexpensive tilt/roll system. In addition, the wide acceptance
© Woodhead Publishing Limited, 2012
356
Concentrating solar power technology
Target (for right side
trough). Strip of cells
Left side trough
primary
Target (for left
side trough)
Right side trough
primary
10.13 Cross section, reflective DSMTS linear concentrator.
angle will allow some of the diffuse radiation to reach the cell, and as a
result, this system should see service in the urban environment.
Limiting the output is, in first place, the performance of the silicon cell
itself. With the high temperature coefficient and rated efficiencies around
17%, the cell is only half as efficient as the best multi-junction cell, which
puts large cost pressures on the optics and mechanical structure. Second,
because of the high flux on the cells, efficiency loss due to temperature rise
occurs, and must be minimized by the use of a very inexpensive but effective
thermal management system on the back side of the receivers. These constraints have been sufficiently extreme to render most attempts at LCPV
unsuccessful until recently.
The mirrors can be fabricated from rolled aluminum sheet. The limited
frequency response of the silicon cells allows this, and is in contrast to most
HCPV systems, where any reflecting surface must be constructed of silver,
for wide bandwidth reflection. The aluminum sheet can be both the mirror
and the support structure. In addition, low concentration allows for more
lenient fabrication tolerances, making forming of the mirror using standard
automotive manufacturing techniques feasible.
Skyline Solar is producing a concentrator based on these techniques, and
has successfully installed at several pilot-sized sites. A fairly new entrant to
the industry, they are combining the inherent advantages of LCPV with a
mature understanding of high volume manufacturing to produce their high
gain solar system. Figure 10.14 is a photograph of one of their systems.
In summary, as can be seen, there are significantly different approaches
to concentrators. The above represent a very incomplete list and is based
on the subset of designs that are furthest along in commercialization. There
are many other ideas and technologies still under development, and a sampling of industrial trends follows.
© Woodhead Publishing Limited, 2012
Concentrating photovoltaic (CPV) systems and applications
357
10.14 Example of a reflective DSMTS concentrator from Skyline Solar
Inc., USA.
10.6
Future trends
The CPV industry is maturing, with several companies deploying welldeveloped products, manufactured on high-volume, automated assembly
lines and complete with high-quality warranties and documentation. The
industry has its first set of qualification standards (IEC62108), and the CPV
Consortium, formed in 2008, now has many members representing the
entire supply and deployment chain.
Interestingly though, there is still a lot of basic technical research to be
done, especially in the areas of optics, high-performance cells, tracking
systems and algorithms, and as a result, the field is still attracting the attention of academic institutions and startup companies worldwide.
The push to commercialization has opened up new fields of research, as
the drive to manufacture, deploy and successfully operate the equipment
has increased. Now that significant installations are being undertaken, new
topics receiving research attention include the interaction of concentrators
and the electricity grid, the effect of weather statistics, and the development
of highly deployable designs.
Therefore, future trends in the industry are beginning to look quite
complex as the research landscape widens, not decreases. A very brief overview of some of the trends follows.
10.6.1
New generation optical systems
Optical systems, at least practical systems for HCPV products, can
still benefit from efficiency and acceptance angle improvements. While
the more common designs described earlier are being honed for
© Woodhead Publishing Limited, 2012
358
Concentrating solar power technology
cost, performance and reliability, there are a few new approaches under
development. As mentioned above, Universidad Polytecnic Madrid has
demonstrated a unique and efficient optical system for use by small (on
the order of 1 mm2) multi-junction cells. This effort has not proceeded past
the prototype stage, however, and has proven difficult to fabricate. If perfected, this approach promises the best acceptance angle for a given
amount of concentration.
Very flat concentrators using TIR have been described (Karp et al., 2010),
and are being commercialized by at least one company. Constructed of
PMMA, they show promise of low cost, albeit with degraded optical efficiency due to long optical path lengths. At present there are no commercially available systems of this type.
10.6.2
Next generation cells
Several companies are marrying the technology of nanostructures (quantum
wells and quantum dots) to the semiconductor junctions of multi-junction
cells. These structures provide the designer some control over the bandgap
of the junction, to tune it for maximum efficiency. Interestingly, nanostructures can widen the bandgap of a junction, rendering it less spectrally
sensitive and possibly allowing a larger energy harvest over the daily spectral variation. No cells with these new structures have yet been
commercialized.
Further, different alloys are being experimented with for the junctions
themselves. The traditional materials used for multi-junction cells were
arrived at by use of relatively low-cost epitaxy techniques, primarily metal
organic chemical vapor deposition, and selection of materials that are lattice-matched. If some of these constraints are lifted, for example moving to
molecular beam epitaxy, more complex alloys can be produced, with interesting PV properties. Several companies are pursuing this path, and are
working through the problems associated with defects and mechanical
strain induced by material groups that have different lattice constants.
The ability to form light traps from nanostructures and/or change the
direction of light so that it traverses a longer path through the cell is being
researched. This will allow more efficient anti-reflection coatings to be
developed, and significantly, will allow these coatings to accept light over a
much larger input angle than the current generation of cells. This will open
the door to higher system concentration ratios.
10.6.3
System level research
CPV systems are interacting with the environment in larger numbers,
prompting research in several fields. Most of this work is new, and aimed
© Woodhead Publishing Limited, 2012
Concentrating photovoltaic (CPV) systems and applications
359
at improving energy harvest, including a better understanding of the effects
of weather and the interaction with the electricity grid.
For example, CPV systems produce a variable output, because of their
dependence on DNI (see Fig. 10.4 and associated explanation above). This
could lead to grid stability difficulties when, for example, intermittent clouds
modulate the output of a large CPV plant (e.g., 50 MW) that contributes
significantly to local generating capacity. A geographically dispersed 50 MW
CPV plant – say ten 5 MW systems positioned along a transmission line –
would mitigate this effect. The question is, what is the optimum spacing,
how geographically sensitive is it and how is the spacing determined?
Another area undergoing scrutiny is that of inverter granularity. An
inverter may be provided per panel, per string, per system, per field subset
and finally, per field. At one extreme, it is reasonably simple to show that if
it were possible to produce an inverter or DC power management device
per cell, energy harvest would be maximized. This is also, though, the most
complex solution, and potentially the least reliable and most expensive. At
the other extreme, a field-level inverter is large, inexpensive and efficient.
Energy harvest suffers, however, from the interaction of a large number of
electrically paralleled strings in which the adverse effects of module performance variability and daily system-to-system shading cannot be mitigated. This tradeoff is under investigation by several power management
companies with a view to optimizing architecture.
There are many other examples of research being done at the system
level that will positively influence the industry in the near future. As a result,
it is expected that the improving trends in performance and cost will continue into the foreseeable future.
10.7
Conclusions
Photovoltaic concentrators are a relative young technology, but have established commercial credentials. While several companies are arguably well
down the commercialization path, there are still many newcomers to the
scene, and there is a visible intellectual churn within the arena. Because of
its youth, the industry enjoys the benefit of potential for improved performance and decreased cost, which is spurring investment, research, and
development.
Concentrators are ideally suited to automotive manufacturing techniques,
since they are predominantly an assembly operation. Automotive capital
equipment is readily available, inexpensive and well tried. Systems are often
assembled from components that have analogs in other industries, and
employ materials that have been in use for many years. Thus, despite their
relatively recent appearance on the photovoltaic scene, if well designed,
they can be robust and reliable.
© Woodhead Publishing Limited, 2012
360
Concentrating solar power technology
These factors have contributed to a rapidly maturing industry that is
capable of increasingly competitive electricity costs as volumes grow and
new ideas are embedded in products. Sustained performance improvements
and the greater adoption of automation will ensure the permanent place of
CPV within the pantheon of solar technologies.
10.8
References and further reading
American Society for Testing and Standards (ASTM) (2008), ‘ASTM G173-03(2008)
Standard tables for reference solar spectral irradiances: direct normal and hemispherical on 37° tilted surface’.
Baxter R (2006), Energy storage: A nontechnical guide, PenWell.
Benitez P, Mohedano R, Miñano J (1997), ‘DSMTS: a novel linear PV concentrator’,
Photovoltaic Specialists Conference 1997, 1145–1148.
Bett A, Dimroth F, Siefer G (2007), ‘Multi-junction concentrator solar cells’, in
Luque A, Hegedus S, Handbook of photovoltaic science and engineering, John
Wiley & Sons.
Gilman P, Blair N, Mehos, M, Christensen C, Janzou S (National Renewable Energy
Laboratory), Cameron C (Sandia National Laboratories) (2008), ‘Solar Advisor
Model User Guide for Version 2.0’, NREL technical report NREL/
TP-670-43704.
Gordon J (2010), ‘Aplanatic optics for solar concentration’, Optics Express, 18 (S1).
Green M (1998), Solar Cells. Operating principles, technology and system applications, Prentice-Hall.
Karp J, Tremplay E, Ford J (2010), ‘Planar micro-optic solar concentrator’, Optics
Express, 18 (2), 1122–1133.
Katz E, Gordon J, Feuermann D (2006), ‘Effects of ultra-high flux and intensity
distribution in multi-junction solar cells’, Progress in photovoltaics: research and
applications, 14, 297–303.
Kennedy C, Terwilliger K, Jorgensen G (2007), ‘Further analysis of accelerated
exposure testing of thin-glass mirror matrix’, NREL report ES2007-36182.
Kurtz S, Geisz J (2010), ‘Multijunction solar cells for conversion of concentrated
sunlight to electricity’, Optics Express, 18 (S1).
Leutz R, Suzuki A (2001), Nonimaging Fresnel Lenses. Design and performance of
solar concentrators, Springer-Verlag.
Lorenzo E (2002), ‘Energy collected and delivered by PV modules’, in Luque A,
Hegedus S, Handbook of photovoltaic science and engineering, John Wiley & Sons.
Luque A (1989), Solar cells and optics for photovoltaic concentration, IOP Publishing Inc.
McDonald M, Barnes C (2008), ‘Spectral optimization of CPV for integrated energy
output’, Proc SPIE, Vol 7046, 704604.
Melia J, Horne S, Klaren A (2010), ‘Rational ratings’, Proc 25th European Photovoltaic Solar Energy Conference.
Miyashita N, Shimizu Y, Okada Y, Institute of Applied Physics, University of Tsukuba
(2007), ‘Effect of increasing nitrogen composition on the performance of GaAs/
GaInNAs heterojunction solar cells’, Proceedings 22nd European Photovoltaic
Solar Energy Conference, 414–418.
© Woodhead Publishing Limited, 2012
Concentrating photovoltaic (CPV) systems and applications
361
Nishikawa W, Horne S (2008), ‘Key advantages of concentrating photovoltaics
(CPV) for lowering levelized cost of electricity (LCOE)’, Proc 23rd European
Photovoltaic Solar Energy Conference, 3765–3768.
Olson J, Kurtz S (National Renewable Energy Laboratories) (1993), Currentmatched high-efficiency, multi-junction monolithic solar cells, US Patent number
5,223,043, issued Jun 29, 1993.
Olson J, Friedman D, Kurtz S (2007), ‘High-efficiency III-V multijunction solar cells’,
in Luque A, Hegedus S, Handbook of photovoltaic science and engineering, John
Wiley & Sons.
Rainhart L, Schimmel W (1974), ‘Effect of outdoor aging on acrylic sheet’, Solar
Energy, 17, 259–264.
Reich-Weiser C, Dornfield DA, Horne S (2008), ‘Environmental assessment and
metrics for solar: case study of SolFocus concentrator systems’, IEEE PV Specialists Conference, San Diego.
Rosenthal A, Lane C (1991), ‘Field test results for the 6MW Carrizo solar photovoltaic power plant’, Solar Cells, 30, 563–571.
Sala G, Luque A (2007), ‘Past experiences and new challenges of PV concentrators’,
in Luque A, Andreev V, Concentrator photovoltaics, Springer.
Sangani C, Solanki C (2007), ‘Experimental evaluation of V-trough (2 suns) PV
concentrator system using commercial PV modules’, Solar Energy Materials and
Solar Cells, 91, 453–459.
Short W, Packey D, Holt T (1995), A manual for the economic evaluation of energy
efficiency and renewable energy technologies, National Renewable Energy Laboratory, March.
Spectrolab (2010), CDO-100-C3MJ Concentrator Solar Cell data sheet. Available
from http://www.spectrolab.com/DataSheets/PV/CPV/CDO-100-C3MJ.pdf.
Swanson R (2002), ‘Photovoltaic concentrators’, in Luque A, Hegedus S, Handbook
of photovoltaic science and engineering, John Wiley & Sons.
Thomalla E, Köpke P, Müller H, Quenzel H (1983), ‘Circumsolar radiation calculated for various atmospheric conditions’, Solar Energy, 30 (6), 575–587.
Vivar M, Herrero R, Martínez-Moreno F, Moretón I, Antón I, Sala G (2008), ‘Effect
of soiling in PV concentrators: mechanisms of light dispersion and real field performance of soiled flat modules and CPV’s’, Proc 23rd Photovoltaic Solar Energy
Conference, Valencia, Spain, 142–145.
Welford W, Winston R (1989), High collection nonimaging optics, Academic Press.
Winston R, Miñano J, Benítez P (2005), Nonimaging optics, Elsevier.
© Woodhead Publishing Limited, 2012
11
Thermal energy storage systems for
concentrating solar power (CSP) plants
W.-D. S T E I N M A N N, German Aerospace Center, Germany
Abstract: The integration of thermal energy storage systems enables
concentrating solar power (CSP) plants to provide dispatchable
electricity. The adaptation of storage systems both to the solar energy
receiver system and the power cycle of the plant is essential. Three
different physical processes can be applied for energy storage: sensible
heat storage in solid or liquid media, latent heat storage using phase
change material and thermochemical energy storage. This chapter gives
an overview of the various technical concepts developed for thermal
energy storage in CSP plants and describes their states of development,
their potentials for use and their performances.
Key words: thermal energy storage, sensible energy storage, molten salt,
steam accumulator, latent heat energy storage, phase change material.
11.1
Introduction: relevance of energy storage for
concentrating solar power (CSP)
This chapter provides a survey of the status and current developments of
storage technology intended for integration into CSP plants. CSP plants
already have an inherent storage capacity in the thermal mass of the
working fluid and of components such as absorbers, heat exchangers and
tubes. Storage units extend this capacity by storing the energy provided by
the solar receiver that is not immediately used by the thermal process of
the CSP plant. During discharge, the storage unit provides heat to the
thermal process and thereby replaces all or part of the solar collector. The
possibility of integrating cost-effective local storage capacity is one of
the most distinct advantages of CSP over other renewable energy technologies. Storage can be integrated in CSP system designs in a manner that
delivers benefits with minimal or zero impact on overall system efficiency
and cost of energy, in distinction to technologies such as photovoltaics,
which must first generate electricity and then add the extra investment and
efficiency loss of a complex independent electrical storage subsystem.
CSP can profit from storage in various ways:
•
Electricity generation can be shifted to periods with high demand thus
increasing the value.
362
© Woodhead Publishing Limited, 2012
Power
Thermal energy storage systems
Compensation of
Storage of cloud transients
excess energy
Charging
Shift of power
production
following demand
Preheating of
components
before sunrise
363
Discharge
Nominal power
demand
for power cycle
Power from
receiver
Time of day
11.1 Various functions of thermal storage in a CSP plant.
•
•
•
The capacity of external backup systems needed to compensate the
mismatch between availability and demand can be reduced; fossil-fired
backup systems have significant investment costs; the low thermal efficiency of these systems diminishes the benefit of electricity generation
from renewable energy sources.
Improved efficiency by avoiding transients in the power cycle due to
clouds.
Reduced start up period by preheating of absorber systems using energy
provided by the storage system (Fig. 11.1).
The economic advantages of storage integration into CSP plants in the
south-western US are described by Sioshansi and Denholm (2010). The
most important conclusion is that generally storage improves the cost efficiency of CSP plants, but the degree of improvement varies over a wide
range depending on the technology and project-specific assumptions made.
This study also addresses the option to shift the power production to periods
with lower ambient temperature thus increasing the efficiency of the power
cycle.
The majority of today’s commercial thermal storage systems used in
industry and solar heating are operated at temperatures below 100°C and
show storage capacities of less than 1 MWth. Storage systems intended for
CSP differ from these systems in two main aspects: CSP and solar process
heat applications demand a temperature range between 120 and 1000°C,
introducing specific requirements regarding corrosion and thermal stability
of materials. Another characteristic of CSP applications is the huge capacity
of storage units. A 50 MWel parabolic trough power plant requires about
1 GWhth storage capacity for a seven-hour operation time. Since the energy
density of thermal storage systems is limited by physical constraints, typical
CSP storage systems require several tens of thousands tons of storage
© Woodhead Publishing Limited, 2012
364
Concentrating solar power technology
material. Due to the large quantities of storage material, the capital costs
of thermal storage systems are usually dominated by their material
costs. Fluctuations in market prices therefore limit the accuracy of cost
estimations.
It is self-evident that a thermal energy storage system aims to minimize
the losses in stored energy subject to cost constraints. However, in comparing systems, the issue of temperature differences and the associated loss of
potential to generate power must also be considered. Every thermal system
must be charged with a heat source at a higher temperature than itself and
then will return heat to a working fluid at a temperature lower than itself.
The efficiency of power generation reduces with lower temperature inputs,
so the presence of the energy storage system may either force lower temperature, lower efficiency power generation, or force higher temperature
operation of solar receivers to compensate, with a consequent increase in
receiver losses.1
Basic storage concepts can be classified into three main groups according
to the physical concept used for heat storage. While today’s commercial
storage systems apply sensible heat storage, the development of direct
steam generation in the absorbers has sparked research activities aiming at
high temperature latent heat storage systems, which have now reached an
advanced status of maturity. The third main group comprises the application
of reversible chemical reactions and sorption processes. Usually, this group
is denoted as chemical energy storage. In the following sections, the various
storage concepts which have been developed within the three main groups
will be described.
11.1.1
Current commercial status of storage technology
Research on storage systems has accompanied the evolution of CSP technology almost from the beginning. Table 11.1 gives a survey of storage
systems integrated into experimental and commercial CSP plants. Today,
many commercial CSP plants already include storage capacity. The development of storage systems for CSP is characterized by a large variety of basic
concepts reflecting the diversity of absorber systems, heat transfer media
and power cycles used in solar thermal power plants. The identification of
the optimal concept for a given application depends on the specific boundary conditions including working fluid, temperature range, storage capacity,
power level and reaction time.
1
In the language of thermodynamics, direct losses of energy are quantified via a ‘thermal’ or
‘first law’ efficiency. The effects of unavoidable temperature drop are quantified with a ‘second
law’ or ‘exergetic’ efficiency.
© Woodhead Publishing Limited, 2012
© Woodhead Publishing Limited, 2012
SEGS-1 (USA)
TSA (Spain)
Solar Two (USA)
PS10 (Spain)
Andasol-1 (Spain)
Solar Tres/Gemasolar
(Spain)
CESA-1 (Spain)
Themis (France)
Nio central receiver
(Japan)
Solar One (USA)
Thermocline with
thermal oil
Two-tank molten salt
Two-tank thermal oil
Two-tank molten salt
Two-tank thermal oil
Packed bed with air
Two-tank molten salt
Steam accumulator
Two-tank molten salt
Two-tank molten salt
Two-tank molten salt,
steam accumulator
Two-tank liquid
sodium
Steam accumulator
Eurelios (Italy)
SSPS (Spain)
Concept
Project
Trough
Tower/heliostat
Tower/heliostat
Tower/heliostat
Trough
Tower/heliostat
Tower/heliostat
Tower/heliostat
Tower/heliostat
Tower/heliostat
Tower/heliostat
Tower/heliostat
Concentrator type
305°C
700°C
565°C
245°C
385°C
565°C
340°C
450°C
300°C
250°C
530°C
430°C
Max temperature
120 MWhth
1 MWhth
110 MWhth
20 MWhth
1,010 MWhth
2,300 MWhth
3 MWhth
40 MWhth
28 MWhth
3.0 MWhth
1.0 MWhth
0.5 MWhth
Thermal capacity
Table 11.1 Survey of selected storage systems integrated in commercial and experimental CSP facilities
1984
1993
1996
2006
2009
2011
1983
1984
1982
1981
1981
1981
Start of operation
366
Concentrating solar power technology
11.2
Sensible energy storage
In sensible heat storage systems, variations of the stored energy ΔQ12 are
dependent on the variation of the mean temperature (T) according to:
T2
ΔQ12 = m ∫ c (T ) dT
T1
where m is the mass and c the mass specific heat capacity. The capacity of
sensible heat storage systems is limited by the available temperature difference T1 − T2 and physically by the specific heat capacity of the storage
material. Using water as an example, since it is the substance with the
highest specific heat capacity per mass of all liquids and solids, the maximum
storage capacity for sensible heat storage systems is in the range of
0.11 kWhth/kg for a temperature difference of 100 K. Compared to chemical
energy sources (e.g. petrol 11.5 kWhth/kg), thermal storage systems require
large masses due to a low storage density.
The development of a sensible heat storage system starts with the selection of the storage material. Systems using liquid storage media can be
distinguished from systems using solid storage media. Another criterion for
classifying storage systems is the concept chosen for transferring solar
energy to the storage material. For direct storage systems, the heat transfer
fluid used for absorbing the solar radiation is also used as storage medium.
Indirect storage systems use a storage medium that is different from the
heat transfer fluid. Figure 11.2 shows a classification of various concepts for
sensible heat storage. These concepts are discussed in detail in the following
sections.
11.2.1
Liquid storage media: two-tank concept
The two-tank storage approach is illustrated in Fig. 11.3. As the name
implies, a liquid medium is cycled between a hot tank and a tank at a lower
temperature which is referred to as the ‘cold tank’ despite being still at
several hundred degrees. Table 11.2 shows various liquid media and properties relevant for thermal energy storage. The simplest way to store sensible
heat in a liquid is to connect inlet and outlet of a heat transfer process to
two separated volumes held at different temperatures. This approach is
widely used for low temperature applications using water as storage medium
in vertical single tanks with a separation of the hot and cold volumes by
thermal stratification. Due to the large volume needed for CSP applications,
two separate storage tanks are usually preferred instead of a single vertical
tank. If a non-pressurized heat transfer fluid (HTF) is used in the absorbers,
the direct storage of this heat transfer fluid is a straightforward solution,
© Woodhead Publishing Limited, 2012
© Woodhead Publishing Limited, 2012
Cold
Molten salt
Hot
Two tank
Cold
Molten salt
Hot
Two tank
Indirect
11.2 Concepts for sensible heat energy storage.
Steam accumulator
Direct
Liquid storage media
Direct
Solid particle
Sensible energy storage
Indirect
Integrated
regenerator
Solid storage media
Cold
Hot
Packed bed
368
Concentrating solar power technology
Heliostat field
290 °C
565 °C
Molten salt loop
Cold salt tank
Hot salt tank
Turbine
G
Steam generator
Condensor
11.3 Simplified scheme of central receiver with two-tank storage
concept.
provided that the HTF is inexpensive. The SEGS I parabolic trough power
plant used direct storage of thermal oil operated between 240°C and 307°C
(Pilkington Solar Int., 2000). Due to the limited maximum temperature,
flammability and the high costs, thermal oil is not considered an attractive
option. Molten salts represent an alternative already known from other
applications (Bohlmann, 1972; Silverman and Engel, 1977). Mixtures of
nitrates have been preferred so far for energy storage in molten salts. The
major drawback of molten salt systems is the allowable operating temperature range, which is limited by the freezing point at the low end and the
onset of thermal decomposition at the high end. Freezing within the storage
tanks must be avoided, since due to the low thermal conductivity, re-melting
is extremely complex. The corrosivity of molten salts increases with temperature; storage tanks exposed to higher temperatures require more
expensive materials. The two-tank concept using molten salt was successfully demonstrated within the Solar Two project, using 1,400 tons of molten
© Woodhead Publishing Limited, 2012
Thermal energy storage systems
369
Table 11.2 Examples for liquid media for sensible heat storage. Note that the
thermo-physical data is indicative for the materials classes and also varies with
temperature
Medium
Specific
heat
capacity
(kJ/kg/K)
dyn.
viscosity
(kg/ms)
798
4.865
800
Density
(kg/m3)
Saturated water
(250°C, 40 bar)
Mineral oil
(<320°C)
Synthetic oil
(<400°C)
Nitrate salt
(220°C <,
<570°C)
Temperature difference =
100 K
Volume spec.
storage
density
(kWhth/m3)
Capacity
specific
media costs
(c/kWhth)
1.06exp-4
100
–
2.4
3.5exp-4
53
15
755
2.4
1.6exp-4
50
60
1,950
1.5
3.4exp-4
81
20
salt between 565 and 290°C. Figure 11.3 shows a simplified scheme of a solar
receiver using molten salt both as working fluid and storage medium. A
detailed description of this storage system is provided by Pacheco (2002).
The direct two-tank concept is also used for the 16-hour storage system of
the Gemasolar project owned by Torresol Energy (Fig. 11.4).
In an indirect storage system, the molten salt (for example) used as
storage medium is different from the HTF used in the absorbers. This
concept is chosen if the specific costs of the HTF are higher than the costs
of the storage medium (e.g. molten salt). The separate loops are connected
by a heat exchanger. The first commercial two-tank molten salt storage unit
integrated into the Andasol-1 plant is an indirect concept using 28,500 tons
of binary nitrate salt operated between 292 and 386°C (Relloso and Delgado,
2009). The basic concept is illustrated in Fig. 11.5.
11.2.2
Liquid storage media: steam accumulator
Liquid water is an attractive storage medium due to its high specific heat
capacity, low cost and compatibility. For temperatures exceeding 100°C,
water must be pressurized to be used as liquid storage medium. Steam
accumulators (Fig. 11.6) provide saturated steam during discharge (Goldstern, 1970). The energy for generating saturated steam is taken from a
pressurized water volume in the saturated liquid state. Since the temperature of the saturated steam depends on the temperature of the liquid water
© Woodhead Publishing Limited, 2012
370
Concentrating solar power technology
11.4 Two-tank storage system of the 17 MWel Gemasolar central
receiver plant. © by Torresol Energy. Reproduced with permission
from Torresol Energy, Spain.
Solar field
Turbine
Hot salt tank
Storage system
290 °C
Cold salt tank
Steam generator
385 °C
Oil/salt heat exchanger
G
Condensor
Synthetic oil loop
11.5 Simplified scheme of a parabolic trough plant using thermal oil
as HTF with an indirect two-tank molten salt storage concept.
© Woodhead Publishing Limited, 2012
Thermal energy storage systems
Steam discharging
Steam charging
Isolated
pressure vessel
371
Steam
Liquid phase
Liquid water
charging/discharging
11.6 Scheme of a steam accumulator.
volume, the pressure of the saturated steam provided by the storage system
decreases during the discharge process. The mass of the saturated liquid
water remaining inside the steam accumulator is large compared to the
mass of the saturated steam provided during the discharge process. Assuming a constant mass mliquid of saturated liquid water inside the steam accumulator, the thermal energy provided during discharge from pressure P1 to
pressure P2 can be estimated to be:
ΔQ12 = mliquid cliquid (Tsat ( P 1) − Tsat ( P 2 ))
where Tsat is the pressure dependent saturation temperature and cliquid is the
mean specific heat capacity of liquid water. Figure 11.7 shows the volume
specific amount of steam provided by a steam accumulator dependent on
initial pressure and pressure drops (Steinmann and Eck, 2006).
Steam accumulators are charged by feeding steam into the liquid volume.
The temperature of the liquid volume is increased by condensation of the
steam. While steam accumulators have fast reaction times, the storage
capacity is usually limited economically by the costs of the pressure vessel.
Since steam accumulators provide steam almost instantly for a short period,
this concept can be used as buffer storage to compensate for short cloud
transients.
Steam accumulators are widely used in process industry in the temperature range between 100 and 200°C. The application of steam accumulators
for large-scale CSP plants was described by Gilli and Beckmann (1976).
Some experimental CSP plants such as Eurelios (Strub et al., 1984) and the
Japanese central receiver plant at Nio (Tani et al., 1986) used steam accumulators. The PS10 central receiver power plant, representing Europe’s first
commercial CSP plant starting operation in 2007, uses steam accumulators
© Woodhead Publishing Limited, 2012
372
Concentrating solar power technology
200
ps
150
100
=
ta
r
80
=
t
10
0
ba
r
ba
40
=
t
ar
p st
ar
0b
=2
0
10
t
ps
rt
50
ps
ta
r
ba
r
r
p sta
Volume specific mass of steam (kg/m3)
250
20
ta
r
t
=
60
ba
r
30
40
50
60
70
80
90
Pressure at end of discharge process (bar)
100
11.7 Volume specific mass of saturated steam provided by steam
accumulator for different initial pressures and pressure drops. Dashed
line indicates example with initial pressure = 100 bar and final
pressure 55 bar; steam accumulator delivers approx. 90 kg saturated
steam per m3 storage volume.
as buffer storage. Four tanks with a total storage capacity of 20 MWh enable
a 50% load operation of the 11 MWel for about 50 minutes. The storage
system is charged with saturated steam at 45 bar provided by the central
receiver.
11.2.3
Solid media storage concepts
The application of solids as storage media is motivated mainly by cost
aspects (Table 11.3). The material costs for concrete, per unit energy stored,
for example, are in the range of 10–20% of the corresponding costs for
molten salt, and maintenance costs are also expected to be lower for solid
media storage systems. Additionally, there are no problems resulting from
freezing, evaporation or leakages. Cost-effective solid storage materials
show low thermal conductivities representing the main challenge for the
implementation of an effective storage concept. Various options have been
suggested to overcome the heat transport limitations of the storage material. The goal of these options is to reduce the path for heat transfer from
the bulk of the storage material to the transfer medium.
© Woodhead Publishing Limited, 2012
Thermal energy storage systems
373
Table 11.3 Examples for solid media for sensible heat storage. Note that data
given is indicative
Material
Density
(kg/m3)
Specific
heat
capacity
(kJ/kg/K)
Thermal
Temperature difference =
conductivity 100 K
(W/mK)
Volume specific Capacity
storage density specific media
costs (c/kWhth)
(kWhth/m3)
Aluminium
Concrete
Cast iron
Rock
2,700
2,200
7,200
2,500
0.945
0.72
0.5
0.8
200
1.5
35
1.0–3.0
70.8
44
100
55
90
5
70
–
Solar field
Turbine
Steam generator
Concrete storage unit
G
Condensor
Synthetic oil loop
11.8 Simplified scheme of parabolic trough plant using thermal oil as
HTF with a concrete storage unit.
11.2.4
Solid media with integrated heat exchanger
For pressurized heat transfer fluids, a parallel pipe tube register is usually
integrated into the storage volume. Various castable storage materials such
as concrete or castable ceramics have been investigated for this approach.
A test unit was installed at the Plataforma Solar de Almería, Spain, and
connected to parabolic trough collectors (Figs 11.8 and 11.9). This system
was designed for a thermal power of 350 kWth and a capacity of 400 kWhth,
operated with thermal oil at a maximum temperature of 390°C. A second
test unit intended for storage cycles of 6–8 hours also having a capacity of
400 kWhth was connected to a test rig to allow investigation of the long-time
behaviour (Fig. 11.10). Operated with thermal oil between 300 and 400°C,
© Woodhead Publishing Limited, 2012
374
Concentrating solar power technology
11.9 Concrete storage module (PSA-Almeria) before installation of
insulation.
11.10 Concrete storage module (400 kWhth) connected to test rig,
before installation of insulation.
more than 300 cycles have been performed without any degradation (Laing
et al., 2009a). The same basic design has also been used for a storage unit
intended for operation with superheated pressurized steam (100 bar) at
maximum temperatures of 500°C (Laing et al., 2009b). The specific investment costs for a storage unit with a capacity of 1100 MWhth are estimated
to be in the range of 34 c/kWhth (Bahl et al., 2009).
Currently, the focus of the ongoing development is on options to reduce
the investment costs. These costs are dominated by the heat exchanger
embedded in the storage volume. Various options to increase the effective
© Woodhead Publishing Limited, 2012
Thermal energy storage systems
375
heat conductivity within the storage material have been investigated. The
homogeneous addition of materials having a high thermal conductivity has
not shown a significant potential for cost reduction. In another approach,
additional heat transfer structures are integrated into the storage material
to reduce the number of tubes needed for the heat exchanger.
11.2.5
Packed bed
In a direct contact storage system, there is no intervening wall between heat
transfer fluid and storage medium. Particles of storage material are packed
into a container, the HTF passes through the particles. Direct contact heat
transfer allows extensive volume specific heat transfer areas. The effective
flow cross section can be large, thus reducing pressure losses, especially for
gaseous HTFs. Storage material and HTF must have the same pressure and
must be compatible. A packed bed storage system is integrated into the
Solar Power Tower Jülich. This experimental central receiver plant uses air
at atmospheric pressure as heat transfer medium. The storage is cycled
between 120 and 680°C and has a storage capacity of almost 9 MWth (Zunft
et al., 2010).
Direct contact storage systems can also be operated with liquid HTFs.
Here, the main aim is to displace expensive liquid storage media by costeffective solids. This approach was chosen at the Solar One central receiver
plant, using a mixture of thermal oil, sand and gravel stored in a single tank.
The hot fluid at the top is separated from the cold fluid at the bottom by
buoyant forces. This thermocline system is charged by feeding hot thermal
oil into the top of the tank. During the discharge process, hot oil is taken
from the top of the tank, pumped through a heat exchanger and returned
to the bottom of the tank. Due to the selected thermal oil the maximum
temperature of heat provided by this system was about 315°C. The thermocline concept was also investigated using molten salt as the HTF (Pacheco
et al., 2002).
11.2.6
Solid particles
Receivers with direct absorption of concentrated solar radiation in solid
particles are considered to be an attractive solution for solar chemistry
applications requiring high temperatures (Siegel and Kolb, 2008). Basically,
this approach can also be chosen as an energy storage system. While the
direct absorption of solar energy in non-pressurized solid storage materials
is an attractive option, various issues have to be addressed before commercial scale CSP applications are possible. The long-term stability of the particles must be ensured, the parasitic load needed for the transport of the
particles must be considered. The piping for the transport of the particles
© Woodhead Publishing Limited, 2012
376
Concentrating solar power technology
will experience significant mechanical loads at high temperatures; heat
transfer from the particles to the working fluid of the power cycle is complex
(Tan and Chen, 2010).
11.3
Latent heat storage concepts
Temperature
µtsaturation
Latent heat storage systems employ the enthalpy change of a substance
passing through a phase change (usually solid to liquid) to store energy. The
most prominent advantage of storage concepts using the latent heat associated with the change of state of the storage material is the option to store
energy within a narrow temperature range close to the phase change temperature. In CSP technology the development of absorbers directly generating steam has sparked interest in latent heat storage systems. Here, the
application of storage concepts using sensible heat storage is usually not
cost effective. This can be seen in Fig. 11.11. If a steam process is used both
for charging and discharging a sensible heat store, the charging steam
system will need to run with a much higher saturation temperature (i.e. at
Condensing steam (charging)
e
PCM
Sa
tu
ra
ti
on
lin
Evaporating steam (discharge)
Temperature
(a)
Specific enthalpy
ge m
ra
ti
Sa
tu
edium
µtsaturation
on
lin
e
Condensing steam (charging)
ra
le sto
ensib
S
Evaporating steam (discharge)
(b)
Specific enthalpy
11.11 Necessary reduction of saturation temperature for a system
using steam as working fluid. Comparison of latent heat storage
concept (a) and sensible heat storage concept (b).
© Woodhead Publishing Limited, 2012
Thermal energy storage systems
377
a much higher pressure) than the discharging steam in order for the heat
transfers to take place. The steam with recovered energy will have considerably reduced power generation potential (exergy) compared to the charging steam. If the latent heat of vaporization for the steam is stored and
released by an appropriately matched latent heat-based energy store, with
the superheat component met by a sensible heat store, then this large exergy
loss can be avoided. The development of a latent heat storage system starts
with the selection of the phase change material (PCM). The temperature
of the phase change should correspond to the specific application. Usually,
this demands that the melting temperature of the storage material is close
to the saturation temperature resulting from the desired steam turbine
operating pressure. The phase change process should be physically and
chemically reversible, i.e. no change of the melting temperature or melting
enthalpy should occur over many cycles. For CSP and solar process heat
application using direct steam generation in the absorbers the range for the
melting temperature for candidate PCMs is between 120 and 340°C. Table
11.4 lists materials showing melting temperatures in this temperature range.
Further relevant physical criteria for storage materials are the specific heat
of fusion and the thermal conductivity.
Metals such as tin or lead can be excluded as PCMs due to cost aspects.
A characteristic of the remaining candidate materials is a low thermal conductivity. Consequently, the development of PCM storage systems requires
the identification of cost-effective heat transfer concepts to overcome the
limitations resulting from poor heat conductivity of the storage material.
The various PCM storage concepts can be distinguished by the approach
to ensuring a sufficient heat transfer between storage material and heat
transfer fluid (Fig. 11.12). These concepts involve the addition of other
materials to the PCM volume. The volume specific costs of these additional
materials should be compared to the volume specific costs of PCM (approx.
1600 c/m3) to estimate the acceptable fraction for additional materials. PCM
storage concepts are usually more complex than sensible heat storage
systems, since the storage material often undergoes a significant volume
change during the phase change. These concepts are discussed in turn in the
following sections.
11.3.1
Phase change material (PCM) concept with
extended heat transfer area
The average distance for heat transfer within the PCM is limited in this
approach. By using finned tubes, the effective surface of the tube is extended.
The aim is to replace expensive pressure pipes by less expensive nonpressurized thermally conductive structures. Finned tubes used in PCM
storage differ from finned tubes applied in heat exchangers in several ways:
© Woodhead Publishing Limited, 2012
© Woodhead Publishing Limited, 2012
KNO3-NaNO2-NaNO3(eu)
KNO3-NaNO3(eu)
Tin
LiNO3
NaNO3
Lead
PCM
142
222
232
252
306
327
Melting
temperature
(°C)
2,000
2,000
7,200
1,800
1,900
11,300
Density
(kg/m3)
0.5
0.5
60
0.6
0.5
35
Thermal
conductivity
(W/mK)
60
100
59
360
175
23
Heat of
fusion
(kJ/kg)
33
55
118
180
96
72
Volume
specific latent
heat (kWhth/m3)
Table 11.4 Examples for PCM with melting temperatures in the temperature range relevant for CSP
65
28
600
60
14
100
Capacity specific
media costs
(c/kWhth)
Thermal energy storage systems
379
Thermal energy systems
using phase change material (PCM)
Composite material
with increased
thermal conductivity
Extended
heat transfer surface
Fins,
sandwich
Capsules
Stiff
Infiltration
Intermediate
heat transfer medium
Compound
Flexible
11.12 Concepts for latent heat energy storage.
Table 11.5 Materials for extended surface heat transfer
Material
Thermal
conductivity (W/mK)
Density
(kg/m3)
Volume specific
material costs (c/m3)
Aluminium
Graphite foil
Carbon steel
Stainless steel
Copper
200
150
30
20
350
2,700
1,000
7,800
7,800
8,800
7,000
9,000
14,000
19,000
35,000
the distance between parallel tubes is larger than that of conventional heat
exchangers, the material of the fins must be corrosion resistant in a PCM (e.g.
molten salt) environment and the fins must be able to withstand thermomechanical stress resulting from the volume variation of the PCM within a
small temperature range during phase changes. The material used for the fins
(Table 11.5) is of specific importance: while steel is used for the pressurized
tubes due to its strength, steel does not represent an optimal choice for the
fins due to high volume specific costs and moderate thermal conductivity.
Alternatively, fins can be made of aluminium or graphite (Fig. 11.13). Both
© Woodhead Publishing Limited, 2012
380
Concentrating solar power technology
11.13 Heat exchanger for PCM storage with high fins made of
aluminium.
materials show comparatively high thermal conductivity and lower volume
specific costs. The temperature range for these fin materials is limited if
nitrates are used as PCM by corrosion resistance (aluminium <400°C, graphite <250°C), but regarding the range of the saturation temperature for water,
aluminium can completely cover the relevant temperature range. The feasibility of the finned tube concept embedded in PCM (sandwich concept) has
been demonstrated in various DLR research projects using fins made of
either graphite or aluminium (Table 11.6; Steinmann and Tamme, 2008). A
PCM storage unit with a capacity of 700 kWhth has been designed for operation with steam at 100 bar using fins made of aluminium (Laing et al., 2010)
(Fig. 11.14). This storage unit was integrated into a test loop connected to the
power plant Litoral of Endesa in Carboneras, Spain.
The macro-encapsulation of PCM is another option for increasing the
heat transfer area. Here, capsules filled with PCM are stacked in a pressure
vessel. Macro-encapsulated low temperature PCMs are commercially available for cooling and low temperature applications. For CSP applications,
capsules with metallic walls must be used (Fig. 11.15). Basically, containers
for PCM can be either thin-walled (flexible) with equal pressure inside and
outside or thick-walled (stiff) with different pressures. For systems using
molten salts, the walls must have a minimum thickness to ensure a sufficient
life expectancy regarding corrosion aspects. Consequently, a design using
flexible containers is not possible for nitrates as the PCM. A significant
drawback of stiff capsules is the necessity to include a gas volume to compensate for the expansion of the PCM during melting (Steinmann and
Tamme, 2008). About 60% of the volume inside the pressure vessel can be
filled by PCM capsules. A laboratory-scale test unit was designed and manufactured by DLR. Cylindrical capsules containing a total mass of 4.5 kg of
NaNO3-KNO3 (eutectic) were stored in a pressure vessel (Fig. 11.15).
© Woodhead Publishing Limited, 2012
© Woodhead Publishing Limited, 2012
200
700
5
ITES-Lab
ITES
100
DISTOR II
PROSPER PLUS
2
15
DISTOR I
Max.
power
(kW)
PROSPER
Identification
700
200 sensible + latent
8
55
7
3.5
Thermal capacity
(kWhth)
NaNO3
KNO3-NaNO2-NaNO3(eu)
NaNO3
KNO3-NaNO3(eu)
KNO3-NaNO2-NaNO3(eu)
KNO3-NaNO3(eu)
PCM
Table 11.6 PCM test storage units developed by DLR, using the sandwich concept
14,000
5,500
140
2,000
400
130
Mass
PCM
(kg)
306
145
306
225
145
225
Melting
temperature
(°C)
Steam
Thermal oil
Thermal oil
Steam
Steam
Thermal oil
HTF
382
Concentrating solar power technology
11.14 PCM storage unit using fins made of aluminium, 700 kWhth
thermal capacity, operated with steam at 100 bar, integrated into a
test loop connected to the power plant Litoral of Endesa in
Carboneras, Spain.
11.15 Pipe segment with containers filled with PCM
(macro-encapsulation).
© Woodhead Publishing Limited, 2012
Thermal energy storage systems
383
Although the feasibility of the concept was proven by experiments, this
approach was not pursued any further due to economic aspects. Regarding
costs, the macro-encapsulation of PCM is not very attractive due to the
limited effective volume share of the storage material and the significant
amount of steel needed for the capsules and the pressure vessel. An additional problem is the necessity to ensure a high quality sealing of the capsules, since contamination of the steam due to leakages must be avoided.
11.3.2
Composite material with increased
thermal conductivity
The effective thermal conductivity can be increased by the homogeneous
addition of a material showing a high thermal conductivity. Highly conductive particles can be dispersed in the PCM. In another approach, PCM is
integrated into matrices made of aluminium or graphite. Both these concepts require the addition of significant amounts of highly conductive materials. Since the contact surface between PCM and additives is large, corrosion
problems increase for nitrates and limit the maximum temperature. Mainly
due to cost aspects, this approach does not seem very promising.
11.3.3
Intermediate heat transfer fluid
An intermediate ‘heat pipe’ system based on the evaporation and condensation of a suitable intermediate heat transfer fluid can be used to transfer
energy between a heat exchanger and the PCM. In this concept, the heat
transfer area of the heat exchanger can be smaller than the outer heat
transfer area of the PCM, and oversizing of the heat exchanger can be
avoided. Since the temperature difference between the steam in the heat
exchanger and the PCM undergoing a phase change should be minimized,
the intermediate heat transfer fluid should also undergo a phase change
between the liquid and the gaseous state. During the charging process,
steam is condensed in the heat exchanger, which is covered by the liquid
phase of the intermediate heat transfer fluid. The energy released during
the condensation is used to evaporate the heat transfer fluid. The saturation
temperature of the intermediate heat transfer fluid is lower than the condensation temperature of the steam but higher than the melting temperature of the PCM. When the gaseous intermediate heat transfer fluid contacts
the surface of the PCM, it condenses and transfers the energy associated
with the phase change to the melting PCM. During the discharge process,
the intermediate heat transfer fluid evaporates at the surface of the PCM
and transfers the heat to the steam by condensation at the surface of the
heat exchanger. Essential for this concept is the identification of the intermediate heat transfer fluid. The temperatures for boiling and condensation
© Woodhead Publishing Limited, 2012
384
Concentrating solar power technology
must be adapted to the temperature range of the steam and the melting
temperature of the PCM. Further details are given by Adinberg and Epstein
(2007).
11.4
Chemical energy storage
Chemical energy storage systems utilize the enthalpy change of a reversible
chemical reaction. The interest in these systems is motivated mainly by the
option to store energy at higher densities than other types of thermal
storage. The possibility of storing the reactants at ambient temperature, so
minimizing thermal losses, is also attractive. Although this potential was
identified early in the evolution of CSP technology (Ervin, 1977; Williams
and Carden, 1978; Brown et al., 1992), chemical energy storage systems are
currently in an earlier stage of maturity, and economic issues and system
aspects demand further investigations. As with all energy storage systems,
the overall energetic and exergetic efficiency of closed loop systems must
be considered.
11.4.1
Reversible chemical reactions
The basic concept of chemical energy storage is to absorb excess heat in an
endothermic reaction. The reaction products are stored separately. During
the discharge process, the reaction products are recombined exothermically
and the heat of reaction can be used. The concept is illustrated schematically in Fig. 11.16.
Catalytic gas–gas reactions represent one group of reactions considered
for energy storage. An example for this group is the CH4 reforming-methanation reaction, which originates from activities aiming at the storage of
heat generated from nuclear energy:
CH 4 (gas) + H 2 O (gas) ↔ CO (gas) + 3 H 2 O (gas)
The endothermic reforming reaction is carried out in the solar receiver at
temperatures between 800 and 1200°C. The products are cooled to ambient
temperature and stored at high pressure. The reaction is reversed in the
methanator system providing heat in the temperature range of 350–700°C.
There is a large body of work addressing the feasibility of this concept,
culminating in a demonstration system using a volumetric receiver installed
on a solar central receiver reaching a power level of 480 kW (Epstein et al.,
1996; Abele et al., 1996). The storage system shows a storage density of
about 45 kWhth/m3. The group at CSIRO in Australia are continuing to
investigate the solar driven endothermic half of the system for the purposes
of ‘open loop’ solar value adding to natural gas (Stein et al., 2009). Another
example is the dissociation of ammonia:
© Woodhead Publishing Limited, 2012
Thermal energy storage systems
Endothermic reactor
A
385
B
Heliostat field
Storage
A
B
Exothermic reactor
A
B
Power generation
G
11.16 Schematic of a reversible chemical energy storage system.
NH 3 ↔ 1/ 2 N 2 + 3/ 2 H 2
which has been extensively investigated by the Solar Thermal Group at the
Australian National University (ANU) (Lovegrove et al., 2004). A receiver
was operated at a power level of 15 kW with solar energy provided from a
dish system. In order to enhance reaction rates in the exothermic ammonia
synthesis reaction, system pressures up to 30 MPa are proposed. At 10 MPa,
the volumetric storage capacity is in the range 40 kWhth/m3. One of the major
advantages of the ammonia-based system is that the heat recovering exothermic reaction is the well-known Haber Bosch process, employed on a major
scale around the world for fertilizer and explosives production. Hence there
is large-scale proven reactor technology already commercially available.
Thermal dissociation of solids and liquids can also be applied for energy
storage. By addition of solar heat to a liquid or solid, a gas is released.
During the discharge process, the synthesis of the dissociation products
provides energy. One example of this kind of reaction is the dehydration/
hydration cycle:
Ca(OH)2 ↔ CaO + H 2 O
The storage capacity of this system is in the range of 300 kWh/m3 (Schaube
et al. 2010). There is a large number of active investigations of solar-driven
© Woodhead Publishing Limited, 2012
386
Concentrating solar power technology
chemical processes aimed at producing fuels and useful chemical products.
These approaches are also forms of energy storage in a general sense. Such
applications are covered in detail in Chapter 20.
11.4.2
Sorption heat storage
A sorption process can be considered to be a chemical reaction system based
on weaker chemical bonds than the covalent bonds encountered in the
systems described above. In a sorption heat storage system, the sorbent is
heated during the charging process and vapour is desorbed from the sorbent.
During discharging, vapour at a lower temperature is adsorbed (solid
sorbent) or absorbed (liquid sorbent) and heat at a higher temperature level
is released. Most research activities on sorption heat storage aim at heating
applications at low temperature ranges. An example for a medium temperature application is the reaction of NaOH and water, which has recently been
considered for seasonal storage of solar heat (Weber and Dorer, 2008). This
chemical reaction system has been demonstrated driving a steam locomotive, the ‘caustic soda locomotive’ described by Riedler (1883).
11.5
Selecting a storage system for a particular
concentrating solar power (CSP) plant
Storage systems intended for application in CSP plants must not only minimize energy seepage to the environment, but avoiding exergy losses is also
critical. This requires the minimization of temperature differences within
the driven heat transfer during charging and discharging. The storage system
must be adapted both to the receiver system of a CSP plant and to the
thermodynamic cycle. The identification of the optimal basic storage concept
represents the first step in selecting a storage system for a specific CSP
application. Due to the early stage of development and the lack of pilotscale demonstrations for chemical energy storage, only sensible heat storage
and latent heat storage are currently considered here. These two concepts
should be regarded as complementary rather than competing. For the
absorber concept using two-phase working fluids (e.g. wet steam), latent
heat storage systems should be preferred due to their capability to store
the energy provided by the condensation of the working fluid within a
narrow temperature range. Sensible heat storage systems should not be
applied here, since the saturation pressure of the steam provided during the
discharge process must be reduced significantly compared to the charging
process, which causes efficiency losses that are not acceptable for CSP
applications. Steam accumulators are the exception to the rule: saturated
or slightly superheated steam is used to increase the sensible energy of the
pressurized water used as storage medium.
© Woodhead Publishing Limited, 2012
Thermal energy storage systems
387
Water, 1 bar
Water, pressurized
Thermal oil
Concrete
Molten salt
Packed bed
0
100
200
300
400
500
600 700 800
Temperature (°C)
11.17 Characteristic temperature range for various sensible heat
storage concepts.
For CSP systems using single-phase working fluids (e.g., thermal oil,
molten salt, air) undergoing significant temperature variations in the
absorber, sensible heat storage is usually preferred. Latent heat storage
does not offer any important advantage here; the higher volume specific
storage density is usually not relevant for CSP plants. If sensible heat
storage is chosen as the basic storage concept, a further criterion for the
identification of the optimal storage concept is the temperature range of
the specific application. Figure 11.17 shows the characteristic application
range for the various concepts based on sensible heat storage. While for
some concepts, the maximum temperature is below the maximum temperature required for some CSP applications, storage concepts using molten salt
must avoid operation near the freezing temperature of the storage material.
Table 11.7 gives an overview of sensible heat storage concepts and their
specific pros and cons.
11.6
Future trends
Today’s commercial scale CSP plants use either two-tank molten salt
storage systems or steam accumulators. This choice is mostly influenced by
the operational experience already existing for these two concepts. While
this approach represents a low-risk solution, these two storage concepts also
show a low potential for further cost reductions. The technical feasibility of
integrating large-scale thermal storage with CSP plants has been proven by
the Andasol and PS10 facilities. Current research activities focus on various
aspects:
•
•
•
further reduction of capital and operating costs
adaptation to advanced power cycles with higher efficiencies at increased
temperature levels
improved operability.
© Woodhead Publishing Limited, 2012
© Woodhead Publishing Limited, 2012
• Cost reduction due to low cost storage
material
• No risk of freezing
• Pressurized working fluid can be used
directly without intermediate heat exchanger
• Low response time
• High volume-specific power
• Substantial operating experience in process
industry
• No heat exchanger necessary
• Suitable for high temperatures
• Low risk potential
• Substantial operating experience in process
industry
• Direct absorption of solar radiation
Solid media
with
embedded
heat
exchanger
Steam
accumulator
Falling particle
Packed bed
with air as
heat transfer
fluid
• Cost reduction possible with partial
substitution of molten salt by low cost filler
material
• Only single tank needed
• Provides heat at constant temperature
during discharge
• Good behaviour at partial charge
• Option for using storage medium as
working fluid in the solar field as well
Two-tank
(molten salt)
Thermocline
(molten salt)
Advantages
Concept
Table 11.7 Summary of sensible heat storage concepts
• Only experience from lab-scale experiences available
• Parasitic loads needed for transportation of large masses
might be critical
• Mechanical loads of heat exchanger for solids at high
temperatures might be critical
Temperature not constant during discharge process
Cost attractive only for small pressures
Large pressurized vessels necessary
Limited storage capacity
Pressure losses might be critical
Combination with solar absorbers using liquid heat transfer
fluids difficult
• Distances between solar receiver and storage must be limited
•
•
•
•
•
•
• Risk of irreversible freezing
• Complex initial filling procedure
• Limited potential for further cost reductions resulting from
technical improvements
• Total costs strongly dependent on costs of storage medium
• Tank volume approx. two times the volume of the storage
material
• Risk of irreversible freezing
• Complex initial filling procedure
• Filler material must be compatible with molten salt
• About 30% of the storage volume cannot be used due to
thermal boundary layer
• Temperature decreases during discharge
• Repair or maintenance of embedded heat exchangers difficult
• Thermomechanical stress between heat exchanger and
storage medium must be considered
Disadvantages
Thermal energy storage systems
389
Storage systems are investigated on two different levels: while system analysis deals with the interaction of storage units with the other components of
a CSP plant, detailed research on various storage concepts aims at cost
reduction by more efficient material usage based on an improved understanding of the heat transfer processes in the storage system.
11.6.1
System analysis
System analysis aims at the identification of requirements for storage
systems to allow for an optimal integration into various kinds of CSP plants.
Based on simulations of the annual performance of a specific CSP application, correlations between storage capacity and cost benefit can be developed. These correlations enable the definition of cost targets for storage
systems. Estimates of acceptable costs for a storage system should also
consider efficiency benefits resulting from the integration of a storage
system. Efficiency benefits can result from an operation strategy which
profits from the possibility to reduce off-design operation of the power
plant by using storage capacity. The development of new operation and
control strategies for CSP plants with integrated storage capacity is essential for exploiting the full potential of storage technology.
11.6.2
Further development of existing storage concepts
With capital costs in the range of 35–50 c/kWhth, two-tank molten salt
systems provide a benchmark for acceptable investment costs for sensible
heat storage concepts. Various aspects should be considered here: costs
should only be compared for systems with the same temperature range,
identical discharge duration and the same reaction time. For storage systems
operated at higher temperatures, higher investment costs are acceptable,
since the heat provided by the storage system is usually used in a power
cycle showing a higher thermal efficiency. While steam accumulators show
high capacity-specific investment costs, these systems are attractive for compensation of fast transients due to their short reaction times. Since the costs
for the pressure vessels become prohibitive at temperatures exceeding
250°C, future applications of steam accumulators are probably limited.
A typical capital cost structure for two-tank molten salt concepts is shown
in Fig. 11.18 (Kelly and Kearney, 2006). The costs are clearly dominated by
the share required for the molten salt; the cost reduction potential for the
other components is considered limited. The estimated maximum cost
reduction for concepts requiring only a single tank for both hot and cold
molten salt volume is only in the range of 10%, since only the less expensive
cold tank can be eliminated. On the other hand, single tank concepts are
usually more complex and require additional effort to separate the hot and
© Woodhead Publishing Limited, 2012
390
Concentrating solar power technology
Miscellaneous
costs
14%
Heat
exchangers
10%
Insulation
4%
Foundation
7%
Salt tanks
16%
Storage
media (salt)
49%
11.18 Capital cost structure for two-tank molten salt storage concept.
Miscellaneous
costs
10%
Insulation
8%
Foundation
5%
Heat transfer
medium
7%
Heat
exchangers
57%
Storage media
(concrete)
13%
11.19 Capital cost structure for concrete storage concept.
cold volumes. The replacement of the binary nitrate salt used currently by
an alternative liquid is one option to reduce costs, but so far no promising
substance has been identified. Substitution of molten salt by a less expensive filler material (thermocline concept) represents a promising option for
cost reduction, but the long-term compatibility of the materials must be
ensured.
Solid media energy storage systems show a significant potential for cost
reduction due to the low costs of the storage material (WANDA, 2006) (Fig.
11.19). Mandatory for the economic success of these systems is the development of cost-effective heat transfer concepts. The cost benefit of solid
© Woodhead Publishing Limited, 2012
Thermal energy storage systems
391
storage material must not be outweighed by the costs of the equipment for
heat transfer, so the acceptable amount of material needed for heat transfer
structures integrated into the storage material is limited. The only latent
heat storage concept to be demonstrated successfully so far in the 100 kW
power range is the sandwich concept using fins made of either aluminium
or graphite.
For the sandwich concept, the cost share of the PCM is currently estimated to be in the range of 20–25%. For the embedded heat exchanger, an
optimized material usage and the integration of manufacturing aspects are
expected to result in significant cost reductions. The improved design of
containers and manifolds is also assumed to offer a further potential to
reduce costs.
While chemical energy storage concepts show a significant theoretical
potential for future cost-effective storage systems, additional research and
development is required to provide the basis for commercial applications.
The long-term reversibility of the reactions must be ensured. The overall
energetic and exergetic efficiency must be evaluated. The energy flow
schemes are usually more complex than for other basic storage concepts
and the integration of energy flows which are not linked to the storage
material is often essential for the efficiency of a chemical storage concept.
Cost estimations must also consider investments for pressure vessels and
heat exchangers, which are often large for CSP applications.
11.7
Conclusion
The plurality of thermal storage technology reflects the diversity of CSP
systems. There is no storage concept that can be identified as the universal
best solution for all applications; the selection depends strongly on the specific requirements regarding heat transfer fluid, temperature range, power
cycle and tariff structure. The pairing of storage concept, solar absorber and
power cycle is essential for successful storage implementation.
Thermal storage technology for CSP applications has made significant
progress both in commercial application and development. For CSP systems
using single-phase heat transfer fluids, two-tank molten salt storage systems
with capacities in the GWhth range have become a proven standard solution.
Since the potential for cost reductions is limited for molten salt systems,
there is still a demand for innovative storage solutions with a better cost
efficiency. The application of cost-effective solid storage materials is a
promising approach for fulfilling this requirement; efficient concepts for the
transfer of thermal energy between solid storage material and the working
fluid are needed.
Efficient storage for CSP systems using two-phase fluids as the heat
transfer medium in the absorbers requires charging and discharging within
© Woodhead Publishing Limited, 2012
392
Concentrating solar power technology
a narrow temperature range, which cannot be done cost efficiently with
sensible heat storage. Latent heat energy storage for CSP applications has
matured in recent years, and pilot systems using phase change materials are
being operated with steam in the 100 bar range providing up to 700 kWth
during discharge. The application of finned heat exchangers embedded in
the phase change storage material has proven to be a concept that provides
the necessary power densities. Ongoing research activities focus on cost
reductions of the heat exchangers and identification of additional storage
materials fulfilling the requirements in terms of costs, compatibility with the
heat exchanger materials and cyclic stability.
11.8
Acknowledgement
The author would like to thank Torresol Energy for providing the picture
of the Gemasolar storage system.
11.9
References
Abele, M. et al. (1996) ‘Test results of a receiver reactor for solar methane reforming
and aspects of further applications of this technology’, Proceedings of the 8th
Solar Paces Conference, June, Cologne.
Adinberg, R. and Epstein, M. (2007) ‘Conception and design of a thermal energy
storage system’, Proceedings of the 5th IASME/WSEAS Int. Conference on
Heat Transfer, Thermal Engineering and Environment, Athens, Greece, 25–27
August.
Bahl, C., Laing, D., Hempel, M. and Stückle, A. (2009) ‘Concrete thermal energy
storage for solar thermal power plants and industrial process heat’, SolarPaces
Conference 2009, Berlin.
Bohlmann, E.G. (1972) ‘Heat transfer salt for high temperature steam generation’,
Oak Ridge National Laboratory report ORNL-TM-3777.
Brown, D., LaMarche, J. and Spanner, G. (1992) ‘Chemical energy storage system
for SEGS solar thermal power plant’, Journal of Solar Energy Engineering, 114(4),
212.
Epstein, M. et al. (1996) ‘Solar experiments with a tubular reformer’, Proceedings
of the 8th Solar Paces Conference, June, Cologne.
Ervin, G. (1977) ‘Solar heat storage using chemical reactions’, Journal of Solid State
Chemistry, 22, 51–61.
Gilli, P.V. and Beckmann, G. (1976) ‘Design and economy of solar plants with integrated thermal energy storage’, Solar Energy, Proceedings of the UNESCO/WMO
Symposium, Geneva.
Goldstern, W. (1970) Steam Storage Installation, Pergamon Press, Oxford.
Kelly, B. and Kearney, D. (2006) ‘Thermal storage commercial plant design study for
2-tank indirect molten salt system’, Final Report, NREL.
Laing, D., Lehmann, D., Fiss, M. and Bahl, C. (2009a) ‘Test results of concrete
thermal energy storage for parabolic trough power plants’, Journal of Solar
Energy Engineering, 131, 041007.
© Woodhead Publishing Limited, 2012
Thermal energy storage systems
393
Laing, D., Bauer, T., Lehmann, D. and Bahl, C. (2009b) ‘Development of a thermal
energy storage system for parabolic trough power plant with direct steam generation’, Proceedings of Energy Sustainability 2009, San Francisco.
Laing, D., Bahl, C. and Fiss, M. (2010) ‘Commisioning of a thermal energy storage
system for direct steam generation’, Proceedings of the SolarPACES 2010 conference, 21–24 September, Perpignan, France.
Lovegrove, K., Luzzi, A., Soldiani, I. and Kreetz, H. (2004) ‘Developing ammonia
based thermochemical energy storage for dish power plants’, Solar Energy, 76,
331–337.
Pacheco, J.E. (2002) ‘Final test and evaluation results from the Solar Two project’,
Sandia report SAND2002-0120.
Pacheco, J.E., Showalter, S.K. and Kolb, W.J. (2002) ‘Development of a molten-salt
thermocline thermal storage system for parabolic trough plants’, Journal of Solar
Energy Engineering, 124, 153–159.
Pilkington Solar International GmbH (2000) ‘Survey of thermal storage for parabolic trough power plants’, NREL/SR-550-27925, DOI 10.2172/765081.
Relloso, S. and Delgado, E. (2009) ‘Experience with molten salt thermal storage in
a commercial parabolic trough plant. Andasol-1 commissioning and operation’,
SolarPaces 2009, 15–18 September, Berlin.
Riedler, A. (1883) ‘Die Honigmannsche Dampflokomotive mit feuerlosem
Natronkessel’, Zeitschrift des Vereines Deutscher Ingenieure, 27, 729–750.
Schaube, F., Woerner, A. and Tamme, R. (2010) ‘High temperature thermo-chemical
heat storage for CSP using gas-solid reactions’, Proceedings of the SolarPACES
2010 conference, 21–24 September, Perpignan, France.
Siegel, N. and Kolb, G. (2008) ‘Design and on sun testing of a solid particle receiver
prototype’, Proceedings of Conference for Energy Sustainability, 10–14 August,
Jacksonville.
Silverman, M.D. and Engel, J.R. (1977) ‘Survey of technology for storage of thermal
energy in heat transfer salt’, Oak Ridge National Laboratory.
Sioshansi R. and Denholm P. (2010) ‘The value of concentrating solar power and
thermal energy storage’, Technical Report NREL-TP-6A2-45833.
Stein, W., Edwards, J., Hinkley, J. and Sattler, C. (2009) ‘Solar thermal steam reforming’, in Garche, J. (ed.), Encyclopedia of Electrochemical Power Sources, Elsevier,
Amsterdam, pp. 300–312.
Steinmann, W.D. and Eck, M. (2006) ‘Buffer storage for direct steam generation’,
Solar Energy, 80, 1277–1282.
Steinmann, W.D. and Tamme, R. (2008) ‘Latent heat storage for solar steam systems’,
Journal for Solar Energy Engineering, 130, 41–45.
Strub, A.S., Gretz, J. and Palz, W. (1984) ‘Eurelios, the 1 MWel experimental solar
thermal electric power plant of the European Community’, Solar R&D in the
European community, Series B, Thermomechanical solar power plants, 1.
Tan, T. and Chen, Y. (2010) ‘Review of study on solid particle solar receivers’, Renewable and Sustainable Energy Reviews, 14, 265–276.
Tani, T., Ikeda, N., Tanaka, T., Marukawa, A., Wakamatsu, S., Horigome, T., Saishoji,
T. and Fukuda, K. (1986) ‘Results and evaluation of the 1 MW solar thermal
electric power plant at Nio’, in Becker M. (ed.), Solar Thermal Central Receiver
Systems, Vol. 1, Springer Verlag, Berlin.
WANDA (2006) ‘Pre-kommerzielle Entwicklung der WESPE Speicher-Technologie
für den Einsatz in ANDASOL Kraftwerken’, final report, Wanda-Project.
© Woodhead Publishing Limited, 2012
394
Concentrating solar power technology
Weber, R. and Dorer, V. (2008) ‘Long-term heat storage with NaOH’, Vacuum, 82,
708–716.
Williams, O.M. and Carden, P.O. (1978) ‘Screening reversible reactions for thermochemical energy transfer’, Solar Energy, 22, 191–193.
Zunft, S., Haenel, M., Krueger, M., Dreissigacker, V., Göhiring, F. and Wahl, E. (2010)
‘Juelich Solar Power Tower – Experimental evaluation of the storage subsystem
and performance calculations’, Proceedings of the SolarPACES 2010 conference,
21–24 September, Perpignan, France.
© Woodhead Publishing Limited, 2012
12
Hybridization of concentrating solar power
(CSP) with fossil fuel power plants
H. G. J I N and H. H O N G,
Chinese Academy of Sciences, China
Abstract: This chapter summarizes developments in solar hybridization
in conventional fossil fuel plants. Drawing on examples of both
experimental and commercial plants, the chapter reviews the potential
of hybrid solar/thermal power systems and outlines the various
arrangements and methods used in hybrid solar/fossil fuel processes.
Several new hybridization systems are also introduced, before future
challenges for technological improvements in hybridization are
discussed.
Key words: hybrid solar/fossil thermal power generation, thermal
hybridization, hybridization principle and method.
12.1
Introduction
The operation of concentrated solar thermal power plants involves a
number of state-of-the-art technologies; however, questions still remain
regarding the installation of systems with a capacity of over 100 MW, and
the ability to bring costs down in order to facilitate widespread use.
A variety of different technologies are considered as methods for improving the solar-to-electric efficiency of solar thermal power systems and of
reducing cost, but most of these technologies have to contend with theoretical and practical limitations. When collected solar thermal energy is used
at a higher temperature, the thermal cycle has high conversion efficiency,
while the solar collectors have reduced efficiency and are expensive to
produce. On the other hand, when collected solar thermal energy is used at
a lower temperature, the solar concentrators are cheaper to produce, but
the efficiency of the thermal cycle is significantly reduced.
Another key issue for solar-only thermal power plants is dispatchability.
The intermittent nature of solar energy sources can be overcome by the use
of some form of energy storage. The recent advances in the field of energy
storage systems are discussed in detail in Chapter 11. Another viable means
of addressing the dispatchability problem is the integration of concentrated
solar thermal power (CSP) in a conventional fossil plant. Hybridization
395
© Woodhead Publishing Limited, 2012
396
Concentrating solar power technology
offers great potential in allowing the cost-effective exploitation of solar
energy on a scale commensurate with energy requirements.
Hybridized power plants can be divided into two categories: those using
thermal integration and those using thermochemical integration. In thermal
integration processes, hybridization uses solar energy to heat water, saturated or superheated steam in combination with fossil fuel combustion. In
thermochemical hybridization, fossil fuels are used as chemical reactants,
while solar energy serves as the process heat to upgrade or decarbonize the
fossil fuel to produce a cleaner fuel. Thermal hybridization is already used
for industrial application, while thermochemical hybridization is still at the
experimental and demonstration stage.
In contrast to solar-only thermal power plants, a solar hybrid plant can
utilize the existing infrastructure of a conventional power station, thereby
reducing the investment in equipment lowering the cost of power production. At the same time, it allows the problem of the intermittent nature of
solar energy to be avoided. In addition, using solar energy in existing fossil
plants goes some way to alleviating fossil fuel shortage and to reducing
greenhouse emissions, especially CO2. Thus, in the short and medium term,
the development of hybridized solar and fossil fuel power plants is a practical means of accelerating the adoption of solar thermal power technology
on a larger scale.
This chapter describes the forms of solar/fossil fuel hybridization involving coal and natural gas, and examines the various methods of integration.
Hybridization technology, integration system design, and equipment are all
discussed, along with a look forward to future promising developments in
solar hybridization technology.
12.2
Solar hybridization approaches
There are a number of different basic approaches to solar/fossil fuel hybridization. The following sections of this chapter address several of these that
have already been used in commercial plants, along with some advanced
concepts that are still under development.
12.2.1
Fossil backup and boosting of solar thermal plants
Many solar thermal power plants use fossil fuel as a source of backup
energy in the absence of sunlight; this is the most common form of hybridization. The nine commercial solar electric generating systems (SEGS) have
a combined capacity of 354 MW and are the most mature and successful of
the solar-hybrid pure Rankine cycles. These systems use parabolic trough
solar collectors and synthetic oil in a collector loop to transfer thermal
energy to Rankine cycle turbines via heat exchangers. Backup gas-fired
© Woodhead Publishing Limited, 2012
Hybridization of CSP with fossil fuel power plants
397
boilers are used when the temperature of the steam is below that required
by the steam turbines.
12.2.2
Solar-aided coal-fired power plants
Solar hybridization with coal-fired steam production is suitable for countries
with large coal resources, such as China and Australia. The solar heat can
be easily integrated into coal-fired power plants, to work in parallel with
the boiler or feedwater heaters. For example, solar heat can be used to
replace the extraction of steam from the turbine to heat the feedwater. This
technology could be a particularly attractive option for repowering older
coal-fired power plants with a capacity of less than 300 MW. One key
advantage of this system is the enhancement of the output of a coal-fired
power plant without the need to oversize the steam turbine.
12.2.3
Integrated solar combined cycle (ISCC) plants
Conventional combined cycle power plants use a gas turbine (Brayton)
cycle in which the fossil fuel is combusted in series with a steam-based
Rankine cycle. This kind of fossil fuel power plant offers the highest conversion efficiency of all the widely used fossil-fired power generation
technologies.
In an ISCC plant, the concentrated solar heat is introduced into the gasfired ‘combined cycle’ power plant where the solar heat replaces or adds to
the exhaust gas from the gas turbine to produce saturated or superheated
steam. ISCC systems seek to add solar steam to the steam cycle of such
plants, with a view to achieving the benefits of both the pure solar input
and the fossil input with the highest efficiency possible.
12.2.4
Advanced systems
Advanced hybridization systems are those in which solar heat and Brayton
cycles are integrated in order to convert solar energy into electricity, in
contrast to the systems mentioned above, in which solar energy is combined
with the conventional Rankine cycle. There are two main categories of
advanced hybrid power plants. In the first, solar heat is introduced to
preheat the compressor discharge air in the gas turbine cycle, while the
second relies on the hybridization of solar heat with decarbonization of the
fossil fuel via decomposition, steam reforming or gasification to produce a
cleaner fuel, also known as a ‘solar fuel’. The hybridization of solar thermochemical sources with fossil fuels is outside the scope of this chapter. The
production of solar fuels is discussed in detail in Chapter 20. Hybridization
via pre-heating compressed air has been demonstrated by the German
© Woodhead Publishing Limited, 2012
398
Concentrating solar power technology
Aerospace Center (DLR) and Plataforma Solar de Almería (PSA), among
others, and is discussed further in Section 12.6.
12.2.5
The role of different concentrators
A hybrid system can be constructed by adding solar collectors to a conventional fossil fuel power plant. This provides only a modest fraction of the
solar energy needed for a large stand-alone fossil-fired plant. The cost effective design of the solar concentrator field is a key aspect, and is closely
related to the establishment of a good match between solar heat, the concentration ratio of the concentrator and the requirements of the power
plant. Usually, the concentrated higher temperature solar heat (above
500°C) requires a dish or tower collector; while the concentrated lower
temperature solar heat (below 500°C) requires a parabolic trough or linear
Fresnel. The following sections briefly introduce these main types of
concentrators.
Parabolic dish
This system involves a parabolic dish-shaped reflector with the receiver
located at the focal point of the dish. Its concentration ratio is about 1,000–
3,000, and the operating temperature of the receiver is approximately 750–
1,000°C [1, 2]. A great deal of the research carried out on dish systems is
centred around their application to Stirling engines; however, it has also
been shown that direct steam generation can be achieved. These systems
could provide steam at up to the highest temperatures and pressures found
in a fossil-fueled plant. Dish systems are presented in detail in Chapter 9.
Solar tower
A solar tower system involves a large heliostat field with a single receiver
mounted on a tall tower positioned at its center. The working substances
used in the receiver can include water/steam, molten salts, liquid sodium
and air. Its concentration ratio is usually in the range of 150–1,500 and the
operating temperature is about 300–2,000°C. Like dishes, tower systems
have the potential to provide steam at the highest pressures and temperatures within a conventional power plant, either directly or via a heat transfer
fluid. These systems are described in detail in Chapter 8.
Parabolic trough
In this system, parabolic trough-shaped mirror reflectors are used to concentrate sunlight onto receiver tubes placed in the trough’s focal line. The
© Woodhead Publishing Limited, 2012
Hybridization of CSP with fossil fuel power plants
399
operating temperature of a trough receiver ranges between 50 and 500°C.
For example, a heat transfer fluid of synthetic thermal oil can be heated to
approximately 400°C [3], while the molten salt that is used as a heat transfer
fluid in, for example, the Archimede project in the south of Italy, can be
heated from 290°C to 550°C [4]. Direct steam production using trough
systems is also at an advanced stage of development. The steam that is
produced either directly or indirectly with troughs will not be at as high a
temperature as that produced from tower or dish systems; however, it can
be applied at various points within a steam cycle. Parabolic trough systems
are dealt with in Chapter 7.
Linear Fresnel reflector (LFR)
In an LFR system, as discussed in Chapter 6, strips of mirrors rotate around
independent parallel axes to reflect sunlight onto a fixed linear receiver [5].
Its reflector array is a line focus similar to that found in a parabolic trough
system. The operating temperature is between 50 and 300°C and the concentration ratio is about 10–40 [6]. For example, in Puerto Errado Thermosolar Power Plants located in Spain, water is heated from 140°C to 270°C
through LFR collector strings [7]. A case study of the application of LFR
systems to a coal-fired power plant is given in Chapter 13.
12.3
Fossil boosting and backup of solar power plants
The best known fossil backup solar power technology is the famous nine
solar electric generating stations (SEGS) plants built between 1984 and
1991, by the American/Israeli Company, Luz International, in the Mojave
desert in California. The SEGS plants started with an initial 14 MW, followed by six plants of 30 MW, finally reaching a capacity of 80 MWe in the
last two units. In total, they provide 354 MW of reliable capacity which can
be dispatched to the Southern California grid [8]. Figure 12.1 shows the
SEGS III–VII plants located at Kramer Junction.
12.3.1
Process integration and design
The flow sheet of a SEGS plant is shown in Fig. 12.2. The heat transfer fluid
(HTF) is heated up to 393°C through the parabolic trough collector field
and returns to a series of heat exchangers, where the fluid is used to generate high-pressure superheated steam at 100 bar and 371°C. After passing
through the HTF side of the solar heat exchangers, the cooled HTF is
recirculated through the solar field. The superheated steam is then fed to a
conventional reheat steam turbine to produce electricity. The exhaust steam
from the turbine is condensed and returned to the heat exchangers and
© Woodhead Publishing Limited, 2012
400
Concentrating solar power technology
12.1 SEGS III–VII plants located at Kramer Junction (A.R. Akradecki;
reproduced under the Creative Commons Attribution-Share Alike 3.0
Unported license).
feedwater pumps to be transformed back into steam. Given sufficient solar
input, the plants can operate at full power using solar energy alone.
A receiver tube for direct steam generation is being developed, manufactured and tested at elevated steam temperatures by SCHOTT [10]. The
receiver tube can withstand a higher operation pressure of up to 120 bar,
compared to the state-of-the-art HTF receiver tubes, which are designed to
work at only 40 bar; similarly, it can tolerate an increase in the operation
temperature from 395°C to 500°C. These advances will lead to significant
improvements in the solar-to-electric efficiency of solar power plants.
12.3.2
Dispatchability
To enable these plants to achieve a rated electric output during periods of
low solar radiation, such as overcast conditions or at night, a backup natural
gas-fired capability can be used to supplement the solar output. This fossil
backup capability allows the SEGS plants to be fully dispatchable. All of
the existing SEGS plants are hybrid solar/natural gas designs that can take
up to 25% of their annual energy from the natural gas plant. Fossil energy
can be used to superheat solar generated steam (SEGS I), from 307°C out
of the solar field to 415°C, resulting in increased efficiency in the Rankine
cycle. When insufficient solar energy is available, fossil energy can also be
used in a separate fossil-fired boiler to generate steam (SEGS II–VII), or
© Woodhead Publishing Limited, 2012
© Woodhead Publishing Limited, 2012
Superheater
Steam
generator
Preheater
Reheater
HPfeedwater
heater
Expansion vessel
Feedwater
pump
Deaerator
HP-
LPfeedwater
heater
LP-
Condensate
pump
Condenser
Generator
~
12.2 Flow diagram of the SEGS plant for pure solar mode (reproduced with permission
from Ref. [9]).
HTF
pump
Solar collector field
402
Concentrating solar power technology
used in an oil heater in parallel with the solar field (SEGS VIII–IX). The
data relating to the SEGS power plants is given in Table 12.1.
12.3.3
Economic effect
Table 12.2 gives a comparison of the economic performance of trough
plants in solar-only operation mode and hybrid operation mode. The cost
information presented is outdated and should only be used to establish how
the various system configurations relate to each other. The cost of power
quoted for the 30 MWe SEGS VI trough plant is 0.17 $/kWh for a solar-only
plant and 0.141 $/kWh for the hybrid plant with 25% fossil backup in 2002.
In plants planned for the near future, the solar-to-electric efficiency is
expected to improve by approximately one-third over the original SEGS
plants, in large part due to new solar receivers and the use of ball-joint
assemblies. Unit capital costs are lower because of the larger plant capacity
and the more efficient solar field, which helps to reduce the size of the
solar field required. The levelized cost of energy (LCOE) is reduced to
0.11 $/kWh for a solar-only plant and 0.096 $/kWh for the hybrid plant.
Overall, hybridization of solar energy with fossil fuel makes the electricity
cost of a solar power plant 12–17% lower than if solar-only mode was
employed.
12.4
Solar-aided coal-fired power plants
In contrast to the fossil backup technology, the hybridization of solar energy
with coal-fired power plants reduces reliance on fossil fuels and allows
outmoded conventional power stations to be updated.
12.4.1
Hybridization process and arrangement
Three different arrangements used in solar-aided coal-fired power plants
[13], are shown in Fig. 12.3.
•
•
Solar-aided with boiler drum: In this hybridization arrangement, the
solar collector field is operated in parallel with the boiler. The solar heat
at around 400°C converts the feedwater from the economizer of the
boiler into saturated steam which is then returned to the drum of the
fossil-fired boiler.
Solar combined with feedwater: In this arrangement, solar energy at
around 300°C acts as a heat source to preheat the feedwater, instead of
using extracted steam from the turbine. The solar field is connected in
parallel with the powerhouse feedwater heaters. The solar heat at 300°C
can raise the temperature of the feedwater to that required for the
© Woodhead Publishing Limited, 2012
© Woodhead Publishing Limited, 2012
Net output
(MW)
80
80
30
30
30
30
30
30
13.8
Plant
name
SEGS IX
SEGS VIII
SEGS VII
SEGS VI
SEGS V
SEGS IV
SEGS III
SEGS II
SEGS I
307/35
316/40
349/40
349/40
349/40
390/100 reheat
390/100 reheat
390/100 reheat
390/100 reheat
Solar field outlet
(°C/bar)
0.083
0.190
0.230
0.230
0.251
0.194
0.188
0.464
0.484
Solar field
area (km2)
31.5
29.4
30.6
30.6
30.6
37.5
37.5
37.6
37.6
Solar turbine
efficiency (%)
Table 12.1 SEGS power plant data from NREL (from reference [11])
35
43
43
43
43
43
43
53
50
Solar field thermal
efficiency (%)
9.3
10.7
10.2
10.2
10.2
12.4
12.3
14
13.6
Solar-toelectric efficiency
3-hour TES
Gas boiler
Gas boiler
Gas boiler
Gas boiler
Gas boiler
Gas boiler
HTF heater
HTF heater
Dispatchability
404
Concentrating solar power technology
Table 12.2 Solar-only and hybrid operation comparison (from reference [12])
Site: Kramer Junction
Plant size, net electric
(MW)
Collector aperture area
(km2)
Thermal storage (h)
Solar-to-electric
efficiency (%)
Plant capacity factor
Capital cost ($/kW)
O&M cost ($/kWh)
Fuel cost ($/kWh)
LCOE (2002$/kWh)
•
30 MW SEGS plant
Near-Term Trough Plant
Solar-only
Solar-only
30
0.188
Hybrid (25%)
30
50
0.188
0.312
Hybrid (25%)
50
0.312
0
10.6
0
10.7
0
13.9
0
14.1
22.2
3008
0.046
0
0.17
30.4
3204
0.034
0.013
0.141
29.2
2745
0.024
0
0.11
39.6
2939
0.018
0.01
0.096
boiler, usually around 220°C. The steam that does not need to be
extracted is further converted in the turbine, and more output work is
therefore generated from the steam turbine compared to the amount
generated before hybridization using the same amount of input coal. In
this way, this conventional coal-fired power plant is updated to a larger
capacity system without incurring substantial costs.
Solar aided with superheater: This arrangement involved the use of solar
heat to produce part of the superheated steam which is then injected
into the steam turbine. The solar collector field is operated in parallel
with the boiler.
Of the above three arrangements, the combination of solar heat with
feedwater is the easiest to implement. This solar-aided technology can
update repowered coal-fired power plants with a capacity of below 300 MW,
giving them greater capacity. Unlike solar-only thermal power plants, this
solar hybridization power technology does not require solar energy storage
equipment. Also, the existing feedwater heaters are operated in parallel
with a solar-driven feedwater heater. In this way, the hybridized power plant
can be guaranteed to be operated at full capacity even in the case of low
solar radiation.
12.4.2
Case study design
The hybridization of solar energy with coal-fired power technology is
suitable for countries rich in coal resources such as China. In northwest
China, a large number of old coal-fired power plants with small capacity
© Woodhead Publishing Limited, 2012
Hybridization of CSP with fossil fuel power plants
405
Turbine
Solar
field
Boiler
Feedwater heater
(a)
Turbine
Boiler
Feedwater heater
Solar field
(b)
Turbine
Solar
field
Boiler
Feedwater heater
(c)
12.3 Schematics of three solar-aided coal-fired processes: (a) boiling
process; (b) feedwater heating process; (c) boiling and feedwater
heating process.
are still in operation, leading to high levels of environmental pollution. The
northwest of China, especially the Xinjiang region, is rich in solar energy
with 2,550–3,500 hours of sunlight per year and a high average solar radiation of 5,430–6,670 MJ/m2/year [14]. Figure 12.4 shows the design flow-sheet
for a hybridization system that could be implemented in a 135 MWe coalfired power plant in the city of Kashi in the Xinjiang region, where there is
an annual average of 3,037 hours of sunshine [15].
In this typical design, both the two high-pressure heaters and four lowpressure heaters in the existing 135 MW coal-fired power plant are still
maintained. These are operated in parallel with the solar feedwater heating
© Woodhead Publishing Limited, 2012
406
Concentrating solar power technology
Generator
Turbine
G
Boiler
230 °C
140 bar
Deaerator
148 °C
145 bar
Condenser
Feedwater
heater
136 °C
7 bar
78 °C
9 bar
Vessel
Parabolic
trough
12.4 A design flow-sheet for the typical 135 MW solar hybrid coalfired power plant in Xinjiang, China.
collector which consists of parabolic trough concentrators, vacuum tube
receivers, and heat exchanger. The concentrated solar heat at around 300°C
substitutes the high-pressure extracted steam to heat the feedwater to
230°C, before it is injected into the economizer of the boiler. When the
direct normal irradiation (DNI) decreases, the total mass flow of the feedwater and the inlet temperature of the regenerative subsystem may be
maintained at the same levels that were employed before repowering.
However, the flow of the feedwater in the solar collector is decreased, while
the flow in the previous feedwater heaters is increased in order to maintain
a constant inlet temperature for the boiler economizer. In this way, this
repowered plant may be operated at fully-rated output during periods of
low solar radiation.
From a design perspective, the overall power output of this hybrid system
increases the capacity of the coal-fired power plant from 135 MW to
152 MW, reducing the coal consumption from 319.75 g to 284.24 g/kWh.
This increased output is a consequence of the use of solar energy. On the
other hand, looking at the conversion of solar energy into electricity, this
increase in capacity of 17 MW is equal to the capacity of an equivalent
solar-only power plant. The annual solar-to-electricity efficiency in this
hybrid plant can reach 15%, about 3 percentage points higher than the
state-of-the-art SEGS VI technology [11].
It is estimated that the typical unit capital cost would be almost 2,007 $/
kW: this is lower than that of the 30 MW SEGS plant in standard configuration, which has costs close to 3,000 $/kW[12]. The preliminary evaluation of
investment needed for this plant is listed in Table 12.3.
It is worth noting that the temperature range of the concentrated solar
heat at 300°C is usually used for heating water, rather than to generate
© Woodhead Publishing Limited, 2012
Hybridization of CSP with fossil fuel power plants
407
Table 12.3 Preliminary evaluation of investment
Total investment cost
Operation and maintenance (O&M) cost
Annual average coefficient of device
investment (CRF)
Annual cost of device investment
Net increased generation power
Increased electricity
Net increased profit
Payback period
Solar-generation cost
Specific investment cost
Million $
Million $
$
Million $
MWe
Million kWh/y
Million $/y
Y
$/kWh
$/kW
34.78
0.70
0.12
4.09
17.08
52
3.70
9.39
0.09
2,007
electricity. Fortunately, hybridization of a solar system with a coal-fired
power plant may allow low-grade solar heat to provide electricity.
12.4.3
Potential of systems in China
By the end of 2009, the total installed capacity of traditional coal-fired
plants in China is estimated to be 599 GW. Small units (∼300 MWe) were
estimated to generate approximately 30.57% or 183 GW [16]. If these conventional plants are converted into solar-aided coal-fired systems, assuming
2,500 h per year of sunlight, it is estimated that coal consumption will be
reduced by 16.2 million tons per year.
In China, several academic institutes and companies are promoting the
industrial application of this kind of solar hybrid plant. For example, the
Institute of Engineering Thermophysics, Chinese Academy of Sciences, and
North China Electric Power University are developing the technology
needed for system integration. Power generation companies, such as Datang,
Huadian, Guodian and others, have begun pre-feasibility studies for hybrid
plants to be built in Northwest China. In the short term, solar hybridization
plants have been designated a high priority technology and their development is supported by the Ministry of Science and Technology and the
National Natural Science Foundation of China. They are acknowledged as
an appropriate and practical option for solar thermal power plants in China.
12.5
Integrated solar combined cycle (ISCC)
power plants
This section will examine integrated solar combined cycle power technology (ISCC). This technology integrates solar energy into the steam cycle of
a combined cycle power plant. The steam generated by solar heat may be
© Woodhead Publishing Limited, 2012
408
Concentrating solar power technology
Solar
steam
generator
Stack
Solar field
Duct
burner
Heat recovery
steam generator
Gas turbine
~
~
Fuel
Condenser
Steam turbine
12.5 Diagram of an ISCC power plant with a single-pressure-reheat
steam cycle ([17], reproduced with permission).
injected into different parts of the heat recovery steam generator (HRSG)
or directly into the steam turbine of the combined cycle. A schematic
diagram of an ISCC plant [17] is shown in Fig. 12.5.
The main feature of this approach to solar hybridization is that additional
steam cycle power is generated without burning any additional fuel; that is,
all the additional power generated in the steam cycle of the combined cycle
is ‘free’ from a fuel perspective. To achieve this, the size of the steam turbine
may be increased relative to the unit that would be used in a pure gas-fired
application. At times of zero solar input, the oversized steam turbine runs
at part load. Solar thermal input can also be used to reduce the fuel consumption of an ISCC plant. In this approach it is the gas turbine that runs
at part load when solar input is available. Note that reducing the fuel consumption of the gas turbine also reduces its power and exhaust energy.
When planning to integrate steam generated by solar energy into a combined cycle, two main questions must be considered: (a) How much solar
energy should be integrated into the combined cycle? and (b) Where is the
best place in the steam cycle to inject the solar generated steam? The best
means of integrating steam that is generated by solar technology is obviously highly dependent upon the steam conditions that can be generated
by that technology. The following sections discuss how various solar technologies can be integrated into combined cycle power plants.
© Woodhead Publishing Limited, 2012
Hybridization of CSP with fossil fuel power plants
409
Steam Turbine
Solar Field
ACC
Fuel
Air
H
P
S
H
R
H
H
P
E
V
H
P
E
C
H
P
E
C
I
P
S
H
I
P
E
V
H
P
E
C
I
P
E
C
L
P
S
H
L
P
E
V
L
T
E
C
Gas Turbine
Heat recovery steam generator
12.6 A medium temperature solar IGCC. (HPSH: high-pressure
superheater; RH: reheater; HPEV: high-pressure evaporator; HPEC:
high-pressure economizer; IPSH: intermediate-pressure superheater;
IPEV: intermediate-pressure evaporator; IPEC: intermediate-pressure
economizer; LPSH: low-pressure superheater; LPEV: low-pressure
evaporator; LTEC: low temperature economizer; ACC: air cooling
condenser) ([18], with permission).
12.5.1
Process integration and design
ISCC technology can be categorized into three types on the basis of its fluid
temperature capability: high temperature: >500°C; medium temperature:
400°C; low temperature: 250–300°C. The medium temperature type has
been most widely implemented, and will be discussed first.
Medium temperature solar technology
The most common medium temperature solar technology makes use of a
parabolic trough. The integration of extra high-pressure saturated steam
into an HRSG is also proposed in integrated gasification combined cycle
(IGCC) plants (Fig. 12.6). The parabolic trough systems can generate steam
up to ∼380°C. Sending out heated feedwater from the HRSG is also very
common in IGCC plants. Note that it is important to take the feedwater
supply to the solar boiler from the proper location in the steam cycle. The
most convenient location for this is from the discharge of the HP feedwater
pump. Similar to what is observed in SEGS plants, the HTF leaving the
steam generator is at ∼290°C. The feedwater temperature should be ∼260°C
to maximize the heating of feedwater in the HRSG and minimize the
heating of feedwater in the solar field.
© Woodhead Publishing Limited, 2012
410
Concentrating solar power technology
High temperature solar technology
Solar tower systems can generate superheated steam at high pressure and
up to 565°C. These conditions potentially allow solar generated superheated steam to be directly admitted into the HP steam line to the steam
turbine. In addition, steam can be reheated in the power tower along
broadly the same lines as in an HRSG. This allows minimal impact on the
HRSG as the superheating and reheating of the solar steam is carried out
in the solar boiler.
Low temperature solar technology
To date, most linear Fresnel systems have been used to generate saturated
steam at up to 270°C/55 bar. This pressure is too low to allow integration into
the HP system of the steam cycle. In this situation there are two options:
either generate saturated steam at ∼30 bar and admit to the cold reheat line,
or generate steam at ∼5 bar and admit it to the LP steam admission line.
12.5.2
Major equipment design
The design mainly involves the integration of the solar field input with the
combined cycle equipment, particularly the HRSG, steam turbine and heat
sink.
Heat recovery steam generator (HRSG)
It is clear that in most cases the steam generated by the solar field needs
to be re-injected into one of the HRSG sections. One major challenge in
the selection process would be the sizing of this particular section of the
HRSG. Both the fossil and solar generated steam could be accommodated;
alternatively, the section could be sized for only the fossil part [19].
Steam turbine
One of the major decisions in the selection of the amount of solar generated
steam (for a given CSP technology) is the sizing of the steam turbine, which
is connected to both the steam mass flow and pressure. The variation in
these two aspects, particularly during night-time or evening operation, will
affect the cycle efficiency. The volumetric flow to the turbine should be kept
close to design value.
Balance of plant (BOP)
The contribution of the solar generated steam also affects the BOP components such as pumps, motor control centers (MCC), piping and cabling.
© Woodhead Publishing Limited, 2012
Hybridization of CSP with fossil fuel power plants
411
If the purpose of the solar generated steam is to augment the power output
of the plant, then the electrical generator must also be able to cope with
the increased capacity. Due to water scarcity the heat sink (condenser
cooling) in these plants is more often an air cooled condenser (ACC) rather
than a conventional wet cooling tower. In either case, the heat rejection
load is increased or modified for the given ambient conditions.
12.5.3
Typical demonstration plant and project
One of the first ISCC plants to be built was Yazd Solar Thermal Power Plant
in Iran [20], which became operational in 2009. The plant comprises two gas
turbine units with nearly 150 MW capacity each, one steam unit with nearly
150 MW capacity, together with a solar thermal unit with 17 MW equivalent
capacity. With a total capacity of 467 MW, the Yazd power plant [21] was
the eighth largest solar power plant in the world at the start of 2010.
In Ain Beni Mathar, Morocco, an ISCC project of 472 MW is being built,
supported by Global Environmental Facility [22], as shown in Fig. 12.7. The
plant includes a parabolic trough solar component of 20 MW (180,000 m2)
with an expected annual net production of 3.538 GWh. The solar output is
estimated at 1.13% of the annual production, representing 40 GWh per
year. The plant started electricity production in May 2011.
Abener has recently constructed a second ISCC power plant in Hassi’Mel,
Algeria. The complex comprises a 130 MW combined cycle, with a gas
turbine power in the order of 80 MW and a 75 MW steam turbine. A
25 MW solar field, requiring a surface of around 180,000 m2 of parabolic
12.7 IGCC plant in Ain Beni Mathar, Morocco (from [22], with
permission from Elsevier).
© Woodhead Publishing Limited, 2012
412
Concentrating solar power technology
mirrors, is the source of non-fossil energy. Nearly $140 million has been
invested in the project, and it is the first privately financed solar thermal
plant in North Africa [23]. The plant started to produce electricity in July
2011.
In Egypt, a plant with a total capacity of 140 MW [24] is currently under
construction. It has a large solar contribution of 30 MW and is supported
by Global Environmental Facility with a $50 million grant. It is due to start
production on December 30, 2012.
In Italy a solar field of 30 MW is being added to an existing power plant
with a 700 MW capacity. The United States is in the process of building an
ISCC plant in Victorville, CA, with three others planned in California and
Florida. In Mexico an ISCC project was approved by Global Environmental
Facility in 2006 and in India a 150 MW ISCC plant [20] is being planned
with a solar contribution of 30 MW.
In 2010, US utility group Florida Power and Light (FPL) opened the first
hybrid solar thermal facility in the US to connect to an existing combined
cycle power plant at the Martin Next Generation Solar Energy Center [25].
The CSP installation with a capacity of 75 MW is the largest of FPL’s solar
facilities and consists of approximately 180,000 mirrors over roughly 500
acres at the existing plant location, which currently produces up to 2.8 GW.
Table 12.4 shows ISCC projects currently at an advanced stage of development [20].
There is a practical issue with ISCC systems: namely that when solar
energy is not available, the steam turbines have to run at part load and thus
with reduced efficiency. In other words, solar steam is only supplied during
some 2,000 of the 6,000–8,000 operating hours of the combined cycle. This
means that the solar share obtainable is less than 10%; as a result ISCC
systems are only considered to have short-term prospects [26].
12.6
Advanced hybridization systems
Advanced hybridization refers to the combination of solar energy with a
gas turbine cycle. There are two main categories: systems that use solar heat
to preheat the compressor discharge air in a gas turbine cycle; and those
that use solar heat to decarbonize fossil fuel for electricity generation. The
latter also involve high-temperature and mid-temperature thermochemical
hybridization processes. Here, we focus on the solar preheating of air and
mid-temperature solar thermochemical hybridization technology.
12.6.1
High-temperature solar air preheating
The concentrated solar power can be used to heat pressurized air in a gas
turbine before it enters the combustion chamber. There is thus no need to
© Woodhead Publishing Limited, 2012
© Woodhead Publishing Limited, 2012
ISCCS
ISCCS
ISCCS
ISCCS
?
ISCCS
ISCCS
ISCCS
ISCCS
ISCCS
Iran, Yazd
Algeria, Hassi R’mel
Morocco, Ain Beni Mathar
Egypt, Kuraymat
Florida Power and Light
US, Victorville, CA
US, Indiantown, FL
Italy, Siracusa
US, Palmdale, CA
Mexico, Sonora State
PT
PT
PT
PT
PT
PT
PT
PT
PT
Technology
Country
Table 12.4 ISCC projects under development
467
150
472
140
?
563
1,125
730
570
500
Capacity
(MWe)
50
75
30
50
30
17
25
20
40
Solar capacity
(MWe)
8.9%
6.7%
4.1%
8.8%
6.0%
3.6%
16.7%
4.2%
28.6%
Solar share
2,200–2,600
–
2,100
2,200–2,600
2,600
2,500
2,300
2,300
2,400
DNI (kWh/m2/y)
Complete
Well advanced
Well advanced
Well advanced
Complete
?
?
?
?
?
Construction status
at December 2011
414
Concentrating solar power technology
use fossil fuels to preheat the air from 300°C to 1200°C, and the solar heat
is converted with all the high thermal efficiency of a modern recuperated
or combined gas turbine cycle.
Solar air preheating has great potential to reduce the costs of solar
thermal power. In addition, the concept could be applied to a wide range
of power levels (from 1 to 100 MW) [27]. Projects involving solar air preheating are discussed below.
Typical projects
The SOLGATE project investigated the concept of combining solar energy
with fossil fuel in a hybrid plant, and showed its feasibility by proving that
a gas turbine could be modified to allow dual operation. Tests were carried
out at different operating temperatures, which resulted in different solar
shares and varying electricity generation costs. In the EU-funded SOLGATE
project demonstrated at the Plataforma Solar de Almería (PSA) [28], pressurized air temperatures of up to 960°C were achieved at the receiver exit.
The aim of the closely related REFOS project is to develop, build and
test modular pressurized volumetric receivers under operating conditions
representative of those encountered when receivers are integrated with gas
turbines. The main focus of the project is the testing of solar air preheating,
accompanied by basic research into materials. The project is led by DLR
and carried out in cooperation with CIEMAT, Spain, and G+H, Germany.
A diagram of a REFOS project principle [27] is shown in Fig. 12.8.
A pressurized volumetric air receiver with a secondary concentrator has
been developed and successfully tested at the PSA, Spain. A number of
receiver modules, each of which consists of a pressurized receiver unit and
a secondary concentrator with a hexagonal entrance aperture (see the
Solar
unit
Combined
cycle plant
Receiver
Gas
G
G
Heliostat
field
Gas
turbine
Steam cycle
12.8 Solar air preheating system (from [27], with permission).
© Woodhead Publishing Limited, 2012
Hybridization of CSP with fossil fuel power plants
Secondary concentrator
Window
Absorber
415
Air inlet
Air outlet
12.9 REFOS Receiver module (from [27], with permission).
REFOS receiver module [27] in Fig. 12.9) were placed on the tower of a
solar plant.
Since pressurized air receivers in solar tower plants can heat compressed
air in a gas turbine to temperatures up to 1,000°C, it is theoretically possible
to achieve a solar share of 40–90%, and annual solar shares of up to 30%
as a base load. Results from a test rig [27], simulating a 30 MWe solar hybrid
gas-turbine plant with a pressurized volumetric air receiver, show that the
thermal solar share is 28.6% for daytime operation and 15% for full-time
operation; the corresponding net solar-to-electricity efficiency is 18.1% and
15.3%, respectively.
Economical potential
The investment required for a hybrid plant depends on the power level of the
plant and on the solar share proposed; the solar share itself is governed by
the maximum exit temperature of the receiver. The current evaluation shows
[27] that for a 30 MWe solar hybrid gas turbine plant with a pressurized volumetric air receiver, the potential solar LCOE is $0.1275–$0.1367/kWh.
If using modern gas turbine systems in recuperation or combined cycle
mode, the conversion efficiencies for solar heat will be increased to over
50%. For a given solar share, this results in a reduced heliostat field size
and in lower overall costs for the solar aspect of the investment, compared
to the investment needed for solar steam generation. Thus, solar gas turbine
systems are expected to have great potential for market success in the
medium term [29].
© Woodhead Publishing Limited, 2012
416
12.6.2
Concentrating solar power technology
Solar thermochemical hybridization plant
Chapter 20 addresses solar fuels in detail. Any solar-derived fuel can be
used in a combined cycle power plant with a suitably modified combustion
system. One of the most obvious strategies in the short term is solar-driven
natural gas reforming. This could be carried out together with combined
cycle power generation which would otherwise burn natural gas directly.
Case study of medium temperature thermochemical hybridization
This section briefly discusses one example of a cost-effective midtemperature solar thermochemical hybridization plant. Figure 12.10 shows
the arrangement of the mid-temperature solar thermochemical hybridization combined cycle using methanol decomposition proposed by the group
in which the authors of this chapter currently work, at the Institute of Engineering Thermophysics, Chinese Academy of Sciences [30]. It is composed
of two main integrated processes: a solar thermochemical process at around
200–300°C and a gas turbine combined cycle with dual-pressure heat recovery steam generation. The solar thermal heat collected at 200–300°C supplies the heat required for methanol decomposition. The solar fuel produced
with syngas (CO and H2) acts as a fuel to drive the combined cycle, which
Syngas
Reactor
Concentrated
radiation
Compressor
Heater
Combustor
Gas
turbine
~
CH3OH
Air
~
Steam
turbine
HRSG
Pump
Condenser
12.10 Schematic of the new solar/methanol combined cycle hybrid
plant.
© Woodhead Publishing Limited, 2012
Hybridization of CSP with fossil fuel power plants
417
then produces electricity. The net solar-to-electric efficiency is expected to
be about 35%, which is related to the energy-level upgrade of mid-temperature solar heat to a high level of chemical energy in the fuel [31].
Key equipment
The parabolic tubular solar receiver/reactor is the most important element
of this mid-temperature solar hybridization technology. A 15 kW midtemperature solar receiver/reactor prototype has been manufactured and
fabricated [32–34], as shown in Fig. 12.11.
(a)
(b)
12.11 Photograph (a) and close-up (b) of the 5 kW solar receiver/
reactor prototype.
© Woodhead Publishing Limited, 2012
418
Concentrating solar power technology
The prototype consists of 30 m2 parabolic trough solar fields, with a concentration ratio of 30, and an axial copper reactor tube enclosed by an
evacuated transparent glass envelope. The tubular receiver/reactor is positioned along the focal line of a one-axis tracking parabolic trough concentrator oriented in an east-west direction. The National Natural Science
Foundation of China has supported this research into the basic scientific
issues surrounding this mid-temperature solar thermodynamic hybridization process, and the related technological developments are supported by
the National High Technology Research and Development Program of the
Ministry of Science and Technology of the People’s Republic of China. The
technology can also be applied in co-generation power, cooling, and heating
using gas turbines, and may represent a viable starting point for the rapid
development of economical solar thermochemical power generation technology in the short term.
12.7
Conclusions and future trends
The hybridization of solar technology with conventional power generation
technology is an extremely promising energy system and is likely to provide
a major share of renewable bulk electricity production in the near future.
It has enormous potential to be economically superior to solar-only plants
using the same field size. Current research predicts that annual thermal
solar shares could be as high as 40–65%, with levelized solar energy costs
much lower than those of solar-only technology.
Currently, almost all existing hybrid solar and fossil fuel plants suffer from
the relative expense of the process. The greatest challenge facing researchers
and engineers today is to work towards a truly standardized primary technology for solar energy use. The areas with most potential for improvements
are in the efficiency of the energy system itself and in the manufacturing
costs of the different components, which must be substantially reduced.
Currently, one of the most realistic approaches is the repowering of existing coal-fired plants using solar thermal energy to replace extracted turbine
steam. The most significant impact will be seen in larger power plants with
a capacity of more than 100 MW, and this could drive the rapid development of the cost-effective implementation of solar electricity, especially for
countries rich in both coal and solar energy resources, such as China,
Australia, and the United States. The key issue in this approach is the integration of concentrator technology and steam turbines with a flexible
system configuration. The basic goal is to develop sufficient, cheap hybrid
solar–thermal power technology with scalable efficiencies of 15–20% in
solar-to-electricity conversion.
With regard to future developments in solar hybridization systems,
emphasis needs to be placed on highly integrated systems, rather than on
© Woodhead Publishing Limited, 2012
Hybridization of CSP with fossil fuel power plants
419
those that are simply annexed to existing plants. Future systems need to
consider resource hybridization, energy conversion hybridization, and
system simplification, as well as co-pollutant control and cascade utilization
of both concentrated solar energy and fossil fuel energy.
12.8
Acknowledgements
This study was supported by the Natural Science Foundation of China (No.
50836005) and the National Basic Research Program of China (973 Program)
(No. 2010 CB 227301).
12.9
References
[1] Stine W., Diver R., ‘A Compendium of Solar Dish/Stirling Technology’. Sandia
National Laboratories, Albuquerque, NM, 1994.
[2] Groenendaal B., ‘Solar Thermal Power Technologies’. Monograph in the framework of the VLEEM Project, 2002.
[3] Lippke F., ‘Simulation of the part-load behavior of a 30 MWe SEGS plant’.
Solar Thermal Technology Department and Sandia National Laboratoires,
1995.
[4] http://www.nrel.gov/csp/solarpaces/project_detail.cfm/projectID=19.
[5] Grena R., Tarquini P., ‘Solar linear Fresnel collector using molten nitrates as
heat transfer fluid’, Journal of Energy, 2011, 36: 1048–1056.
[6] Singh P., Ganesan S., Yadava G., ‘Performance study of a linear Fresnel concentrating solar device’, Journal of Renewable Energy, 1999, 18: 409–416.
[7] http://www.nrel.gov/csp/solarpaces/project_detail.cfm/projectID=46.
[8] Brakmann G., Aringhoff R., Geyer M., Teske S., ‘Concentrated solar thermal
power now: Exploiting the heat from the sun to combat climate change,
Greenpeace/ESTIA/Solar Paces, 2005.
[9] Stuetzle T., Blair N., Mitchell J.W., Beckman W.A., ‘Automatic control of a
30 MWe SEGS VI parabolic trough plant’. Solar Energy Laboratory, University of Wisconsin-Madison, 2001.
[10] Eck M., Eickhoff M., Feldhoff J., et al., ‘Direct steam generation in parabolic
troughs at 500°C – First results of the Real-Diss project’. Concentrating Solar
Power and Chemical Energy Systems. SolarPaces, 2011, Granada, Spain.
[11] National Renewable Energy Laboratory, ‘TroughNet: Solar parabolic power
network’, 2008.
[12] Price H., Kearney D., ‘Reducing the cost of energy from parabolic trough solar
power plants’. National Renewable Energy Laboratory. International Solar
Energy Conference, Hawaii, 2003.
[13] Odeh S., ‘Unified model of solar thermal electric generation systems’, Journal
of Renewable Energy, 2003, 28: 755–767.
[14] http://scitech.people.com.cn/GB/6772011.html.
[15] Hong H., Zhao Y., Jin H., ‘Proposed partial repowering of a coal-fired power
plant using low-grade solar thermal energy’, International Journal of Thermodynamics, 2011, 14(1): 21–28.
[16] http://www.ndrc.gov.cn/jjxsfx/t20100713_360615.htm.
© Woodhead Publishing Limited, 2012
420
Concentrating solar power technology
[17] Dersch J., Geyer M., Herrmann U., et al., ‘Trough integration into power plants
– a study on the performance and economy of integrated solar combined cycle
systems’. Journal of Energy, 2004, 29: 947–959.
[18] Anon., Renewable energy: The future of power generation, Engineering
Review, September 2009 http://www.engrreview.com/Editorial_pages/2009/
sep_09/power_technofocu-07.html (accessed December 2011).
[19] Kelly B., Herrmann U., Hale M., ‘Optimization studies for integrated solar
combined cycle systems’. Proceedings of Solar Forum 2001.
[20] http://www.zawya.com/story.cfm/sidZAWYA20101220044922.
[21] Siemon B., Alice D., ‘Defining the techno-economic optimal configuration of
hybrid solar plants’. Universiteit Gent Faculteit Economie en Bedrijfskunde
Academiejaar, 2008–2009.
[22] http://www.nrel.gov/csp/solarpaces/project_detail.cfm/projectID=43.
[23] http://www.nrel.gov/csp/solarpaces/project_detail.cfm/projectID=44.
[24] http://www.nrel.gov/csp/solarpaces/project_detail.cfm/projectID=65.
[25] http://www.fpl.com/environment/solar/martin.shtml.
[26] Geyer M., ‘International Market Introduction of Concentrated Solar
Power – Policies and Benefits’. IEA SolarPACES Implementing Agreement.
Proceedings of ISES Solar World Congress 2007: Solar Energy and Human
Settlement.
[27] Buck R., Bräuning T., Denk T., et al., ‘Solar-hybrid gas turbine-based power
tower systems (REFOS)’, Journal of Solar Energy Engineering, Transactions
of the ASME, 2002, 124: 2–9.
[28] Garcia P., Ferriere A., Flamant G., et al., ‘Solar field efficiency and electricity
generation estimations for a hybrid solar gas turbine project in France’,
Journal of Solar Energy Engineering, Transactions of the ASME, 2008, 130:
0145021–0145023.
[29] Schwarzbözl P., Buck R., Sugarmen C., et al., ‘Solar gas turbine systems: design,
cost and perspectives’, Journal of Solar Energy, 2006, 80: 1231–1240.
[30] Hong H., Jin H., Ji J., et al., ‘Solar thermal power cycle with integration of
methanol decomposition and middle-temperature solar thermal energy’,
Journal of Solar Energy, 2005, 78(1): 49–58.
[31] Hong H., Liu Q., Jin H., ‘Solar hydrogen production integrating low-grade solar
thermal energy and methanol steam reforming’, Journal of Energy Resources
Technology, Transactions of The ASME, 2009, 131(1): 0126011–01260110.
[32] Jin H., Sui J., Hong H., et al., ‘Prototype of middle-temperature solar receiver/
reactor with parabolic trough concentrator’, Journal of Solar Energy Engineering, Transactions of The ASME, 2007, 129(4): 378–381.
[33] Liu Q., Hong H., Yuan J., et al., ‘Experimental investigation of hydrogen production integrated methanol steam reforming with middle-temperature solar
thermal energy’, Journal of Applied Energy, 2009, 86(2): 155–162.
[34] Hong H., Liu Q., Jin H., ‘Operational performance of the development of a 15
kw parabolic trough mid-temperature solar receiver/reactor for hydrogen
production’, Journal of Applied Energy, 2012, 90(1): 137–141.
© Woodhead Publishing Limited, 2012
13
Integrating a Fresnel solar boiler into an
existing coal-fired power plant: a case study
R. M I L L A N, J. D E L A L A I N G,
E. BAU T I S TA, M. R O JA S and
F. G Ö R L I C H, Solar Power Group GmbH, Germany
Abstract: A Fresnel solar boiler is used to investigate the influence of a
superheated-steam generating solar add-on on the overall performance
of a 350 MW coal-fired power plant in two promising places for this type
of technology. The solar boiler supplies steam to the preheating stages
of the power block in order to reduce the steam extraction from the
turbine. The focus of the investigation is the evaluation of the potential
of using not just a single but several points of supply for solar steam.
Different combinations of solar steam management are analyzed
concerning their thermodynamic performance, and potential to save
fossil fuel and thereby reduce the carbon footprint of power generation.
Key words: coal-fired power plant, Fresnel solar boiler, solar add-on,
feedwater heating, internal rate of return (IRR), electricity production
cost, emission reduction.
13.1
Introduction
Coal-fired power plants produce electricity using a steam turbine. Retrofitting existing coal-fired power plants is a low-cost option for solar thermal
systems, as most of the existing system can be used. Solar add-ons can be a
way to simultaneously expand the use of solar energy and trim carbon
footprints [1]. The implementation of a solar system into the power plant
does not affect the steam cycle or the electrical efficiency during the periods
of low irradiation, since the fuel supply can be adjusted (within operating
limits) according to the thermal energy available from the solar system.
The integration of a concentrating solar power (CSP) plant into an existing steam-based power plant is readily achieved with a direct steam generating system. In that way it can be directly connected to the steam cycle
and does not require additional heat transfer media and heat exchangers.
Any of the CSP technologies could be applied to this end, however this
chapter presents a case study based on the use of linear Fresnel technology.
Fresnel systems are linear concentrators that use an array of parallel long
mirrors moving to maintain a linear focus on a fixed linear receiver mounted
on towers. Fresnel systems are discussed in detail in Chapter 6. The linear
421
© Woodhead Publishing Limited, 2012
422
Concentrating solar power technology
Fresnel system offers the ability to conveniently generate direct steam,
either saturated or superheated.
The aim of this case study is to demonstrate and assess the potential fuel
and cost savings that can be achieved by retrofitting a linear Fresnel system
in an existing coal-fired power plant.
13.2
Description of options considered as variables
selected for the case study
13.2.1
Coal-fired power plant
A regenerative steam power cycle, with a six-stage feedwater heating system
(five close feedwater heaters and one open feedwater heater and reheat),
was used as the basis for this case study; see the process flow diagram (PFD)
in Fig. 13.1. This simplified PFD represents a typical design of coal-fired
power plants installed in many locations. High pressure (HP), intermediate
pressure (IP) and low pressure (LP) turbine stages are mounted on a single
shaft coupled to the generator. Steam is reheated in the boiler following
expansion in the HP turbine, before returning to the IP turbine, which serves
to prevent excessive condensation of liquid in the LP turbine. In a regenerative cycle, the feedwater is heated progressively in a series of feedwater
heaters (FWH) which involve heat exchange with steam bled off from the
turbines at various points in such manner that in each case the steam used is
Generator
HP
IP
G
LP
Condenser
Cold-reheat line
b
a
Coal-fired
boiler
c
FWH #1
c
FWH #2
c
FWH #3
FWH #4
FWH #5
FWH #6
13.1 Process flow diagram of the coal-fired power plant with possible
solar steam injection points marked a, b and c.
© Woodhead Publishing Limited, 2012
Integrating a Fresnel solar boiler
423
of sufficient temperature to drive the necessary heat transfer. FWH #3 is an
open feedwater heater that mixes steam directly into the feedwater.
The chosen electric capacity of the coal-fired power plant is 350 MW. The
temperature of the water/steam working fluid varies from around 50°C up
to over 500°C, giving a wide temperature range for the possible integration
of the solar system.
13.2.2
Modes of operation when integrating solar steam
The boundary conditions of a retrofitting concept are set according to the
tolerance of the power plant’s components. In ‘augmentation’ mode, the
purpose is to increase the production of electricity by moving the operation
of the turbine-generator set beyond its nominal operating point. Most
systems can typically work at up to 112% of nominal capacity. The boiler
continues operating at nominal full load. If the plant ‘fuel conversion efficiency’ is defined as the ratio between the electricity produced and fossil
primary energy, then the final effect is increased plant efficiency as more
electricity is produced with the same amount of fuel burnt. In ‘fuel-saving’
mode, the boiler operates below its full load point. The solar steam is added
to the main flow in such a way that the turbine receives enough steam to
continue operating at its nominal point. In this case the net effect is that
the same electricity is produced with less fuel.
13.2.3
Solar steam insertion points
Main steam line
There are several considerations regarding the injection point (point ‘a’ in
Fig. 13.1). The most simplistically obvious option is to add the solar steam
to the main steam line. However, this is technically challenging, as the steam
quality has to be assured to avoid condensation inside the turbine that
might damage it. Another important point is the operational temperature
of the turbine; state-of-the-art linear concentrator solar thermal systems can
reach up to a maximum of 450°C superheated steam whereas the operational temperature of the common turbines used today is around 550°C.
This implies the use of additional superheaters to boost the temperature of
the solar steam. Otherwise, the decrease in efficiency due to lower steam
temperature can override the desired efficiency increase effect.
Cold reheat line
In this case solar steam would be injected into the exit steam flow from the
HP turbine, prior to reheat in the boiler (point ‘b’ in Fig. 13.1). For solar
© Woodhead Publishing Limited, 2012
424
Concentrating solar power technology
boilers with the capacity to produce superheated steam, this method of
insertion provides significant additional benefits, due to the net increase in
efficiency obtained as a result of the lower energy requirement of the steam
to reach the turbine operating conditions. This effect is directly related to
the solar boiler temperature: the higher, the better.
Feedwater heater
Another suitable option is to replace or reduce the steam bled from the
turbines for the feedwater heater with the solar steam (points ‘c’ in Fig.
13.1). Usually, to increase efficiency, the turbine is bled to allow preheating
of the water entering the boiler. The same effect is obtained if solar energy
is used to preheat the water, allowing the turbine to extract more power
from the inlet steam. This has the advantage that it is one of the least invasive methods of retrofitting, because it avoids complex boiler integration
issues [2].
Due to the benefits given above, this integration concept is used in the
case study. The high pressure and intermediate pressure bleed extractions
from the turbine are replaced by the solar steam generated and the fuelsaving mode is applied. Although different types of heat transfer fluids can
be used in a Fresnel system, the use of water/steam, known as ‘direct steam
generation’, is less risky, more efficient and requires less investment.
If the goal is simply to heat feedwater, the option of directly passing the
feedwater through the solar boiler or else using any other sort of heat
transfer fluid in heat exchangers for the purpose is valid. The disadvantages
of not using water/vapour as heat transfer fluid are:
•
•
•
addition of heat exchangers → more expensive
decrease of efficiency due to the addition of heat exchangers
fluids, like salt or thermoil, are not environmental friendly and are
expensive.
13.2.4
Solar steam generation
In principle, all of the CSP technologies, trough, tower or Fresnel could be
used as solar add-ons to existing power stations. With these collector types,
either direct steam generation or a range of heat transfer fluids could be
used. The technology selected for this case study is superheated steam –
Fresnel solar boiler, since it has been demonstrated that it can provide the
highest steam conditions and matches the bleed extractions from the turbine.
The Fresnel solar boiler considered, based on the systems produced by
the German company Solar Power Group GmbH, consists of 120 collector
modules. Each collector module has a mirror surface of about 1,400 m2
distributed in 24 rows that concentrate the light onto a horizontal pipe,
© Woodhead Publishing Limited, 2012
Integrating a Fresnel solar boiler
Preheating
section
Evaporation
section
425
Superheating
section
Steam
drum
13.2 Different sections of the Fresnel solar boiler.
located 8 m above, through which water/steam circulates. Each row is 96 m
long and comprises 16 mirrors.
The solar collector field is divided into three sections, analogous to a
regular fossil-fired boiler. In the preheating section, the feedwater is heated
up to a temperature close to the evaporation point at the operating pressure
and then sent to the steam drum. In the evaporation section, the feedwater
coming from the steam drum is gradually evaporated. This section is
designed to have a certain degree of steam wetness at its outlet, meaning
that part of the flow is still in liquid phase. To ensure dry steam conditions
at the entrance of the superheating section, the wet steam coming out of
the evaporation section is separated in the steam drum. The temperature
of the steam is further increased in the superheating section up to the
desired outlet condition. The arrangement of the three different sections of
the solar field is shown in Fig. 13.2.
13.2.5
Integration of the Fresnel solar boiler into the
coal-fired power plant
The Fresnel solar boiler is integrated into the existing coal-fired power plant
by installing it in parallel to the existing high and intermediate pressure
(HP/IP) feedwater heater steam bleed lines (see Fig. 13.3). This arrangement allows high flexibility since the maximum solar energy input can be
© Woodhead Publishing Limited, 2012
426
Concentrating solar power technology
Generator
HP
IP
G
LP
Condenser
Coal-fired
boiler
FWH #1
FWH #2
FWH #3
FWH #4
FWH #5
FWH #6
Fresnel solar
boiler
13.3 Process flow diagram showing the integration of the Fresnel
solar boiler into the coal-fired power plant.
used and low solar heat input can be compensated independently of the
required load demand.
The economic impact and the electrical output improvement of reducing
the HP/IP extraction steam is much higher than for the LP extraction steam
[3]. For this reason, the solar system is designed to replace the bleed extractions from the steam turbine for the HP and IP feedwater heaters, FWH #1
and FWH #2 respectively. The open feedwater (FWH #3) supplies the solar
field with the required feedwater. The amount of feedwater that is pumped
to the solar field depends not only on the feedwater heater demands, but also
on the availability of solar irradiation. The working temperature and pressure of the solar field are given by the operating conditions of the FWH #1,
thus the steam flow is throttled before entering into the FWH #2 and #3.
Regarding the distribution of the solar steam to the preheating system,
two cases have been analysed: one that gives priority to the FWH #1 demand
and the other that supplies the same amount of heat to FWH #1, #2 and #3.
Case A: Steam flow to feedwater heaters distributed by priority
The steam coming out of the superheating section is distributed to the
feedwater heaters according to a given priority, with FWH #1 having the
© Woodhead Publishing Limited, 2012
Integrating a Fresnel solar boiler
427
Table 13.1 Conditions at locations selected for the study [4]
Location Average ambient Average
Available sun Annual direct
temperaturea (°C) relative
hours per year normal irradiation
humiditya (%) (h)
(kWh/m2)
Site 1
32
17
2,480
2,300
Site 2
27
44
2,280
1,950
a
Average ambient temperature and relative humidity during the sun hours.
highest. First, the demand of FWH #1 is covered; any surplus is fed into
FWH #2. In case the steam generated by the solar field exceeds the requirement of FWH #1 and #2, FWH #3 receives the surplus steam produced. The
bleed extractions from the turbine ensure that sufficient steam is provided
to the feedwater heaters regardless of the output of the solar field.
Case B: Split steam flow to the feedwater heaters
In this case, a steam header splits the outlet steam of the solar field into
three streams with the same steam flow rate. The outlets are connected to
FWH #1, #2 and #3. The three outlet flows of the header have the same
composition and specific enthalpy as the header’s inlet (but as noted above,
steam is throttled to adjust its pressure before passing to FWH #2 or #3).
As in case A, the bleed extractions from the turbine ensure that sufficient
steam is provided to the feedwater heaters regardless of the output of the
solar field.
13.2.6
Power plant locations
Suitable locations for concentrating solar power plants are especially arid
and semi-arid regions where the annual direct normal irradiation is above
1,700 kWh/m2. For the purposes of this study, two globally representative
specific locations in high solar areas were used for the analysis. The solar
irradiation of the selected locations is shown in Table 13.1.
13.3
13.3.1
Assessment of the solar add-on concept
Technical assessment
The technical assessment comprises a comparison regarding the thermodynamic performance, fossil fuel saving and CO2 emission avoidance of the
two different cases in the two proposed locations.
© Woodhead Publishing Limited, 2012
428
Concentrating solar power technology
Solar add-ons can either enhance the plant net output or reduce the fuel
consumption, as discussed previously. Since the base case is a hypothetical
existing plant and it is uncertain how much additional steam the steam
turbine and generator could handle, the target is to reduce the fuel consumption and not to increase the electricity production. Thus, the fuel
saving mode is used. The solar field is designed in such a way that its peak
production does not exceed the heat demand of the HP/IP feedwater
heaters at nominal load conditions of the power plant.
Power plants, such as the one presented in this paper, often follow daily
load fluctuation from full load down to the minimum turbine load. In order
to simplify the simulations, the plant is assumed to be a base load plant with
the turbine running at full load continuously.
For the evaluation of the hybrid plant performance, the Thermoflex software [5] was used. Thermoflex is a simulation environment of steady state
of power plants. It is a modular program which offers the opportunity to
design and simulate combined cycles, conventional steam power plants, and
thermal power systems, among others. Its extensive library of components
also includes the elements required to simulate solar thermal power systems.
The development of a model is carried out in two steps. In the first step,
‘design mode’, the thermodynamic design criteria (working parameters) are
determined. In second step, ‘offdesign mode’, the size of the components
and the logic of their controls are established. Once the model is in offdesign, the performance of the system at different operating conditions,
such as the weather data, can be evaluated.
Based on the hourly resolution of the meteorological data supplied by
METEONORM [4] – DNI, sun position, ambient temperature, atmospheric
pressure and relative humidity – the annual performance of solar add-ons
to the coal-fired power plant was assessed by calculating steady state for
each hour of the year. The number of collector modules and thereby the
available mirror surface were kept constant for all simulations. The same
applied for the base model of the conventional coal-fired power plant. The
simulations also take into account the geometry of the solar field, position
of the sun, irradiation level, ambient temperature and relative humidity. The
solar steam production varies according to the feedwater heater demands
and solar irradiation. The electricity production of 350 MW was regulated
by adjusting the coal consumption of the boiler. The inputs, outputs and
design parameters for the simulations are summarized in Table 13.2.
Figure 13.4 shows the amount of thermal energy generated by the solar
field in sites 1 and 2.
The type of steam extractions used is uncontrolled or non-automatic
extraction, the pressure and the flow vary as a function of the load [6,
p. 224]. Therefore, slight variations in the demand of the feedwater heaters
can be seen in Fig. 13.4, which is acceptable.
© Woodhead Publishing Limited, 2012
Integrating a Fresnel solar boiler
429
Table 13.2 Inputs, outputs and design parameters for the simulations
Fresnel
solar
boiler
Power plant
Inputs (METEONORM
data)
Design
parameters
Outputs
→ DNI
→ Solar azimuth
angle
→ Solar zenith angle
→ Ambient
temperature
→ Ambient
temperature
→ Relative humidity
→ Number of
modules (120)
→ Solar steam mass
flow
→ Electricity
production
(350 MW)
→ Fuel mass flow
→ Steam mass flow
of the bleed
extraction 1, 2 & 3
As shown in Fig. 13.4, in case A, the heat demand of the FWH #1 is
covered for about 1,500 hours completely by the solar boiler, while for
FWH #2 approximately 50 full-load hours are covered, in both locations.
The amount of solar energy provided to the FWH #1 (a) is higher than to
the FWH #2 (b) and #3 (c). Whereas in case B, none of the feedwater
heater demands is fully covered by the solar system, not even at the highest
intensity, where only ∼70% of the demand of FWH #1, #2 and #3 is covered.
At lower irradiation levels, the solar energy provided to the three FWHs
diminishes proportionally.
Solar add-on results in a relative plant coal conversion efficiency
(expressed as electricity generation per thermal input from coal combustion) improvement up to 1.6% and 1.5% at sites 1 and 2, respectively. The
increment in efficiency implies lower fuel consumption and therefore higher
CO2 emission avoidance. In both regions, case A has a higher impact on the
overall performance of the plant than case B, since the HP bleed extraction
from the turbine has been totally substituted by the solar system (see
Fig. 13.5).
13.3.2
Economic assessment
Add-ons based on solar direct steam generation have an impact on generation costs, among others, due to the associated fuel savings derived from
their implementation. The expected effect on the electricity production cost
depends on the value of the equivalent fuel savings. The analysis is based
on the contribution of the solar field to electricity generation. Aside from
fuel, generation costs include the amortized costs of the solar collectors and
their operation. Revenues assumed from electricity sales are based on the
calculated solar contribution to electricity production.
© Woodhead Publishing Limited, 2012
© Woodhead Publishing Limited, 2012
Thermal power (MW)
Thermal power (MW)
0
20,000
40,000
60,000
80,000
100,000
120,000
140,000
160,000
0
20,000
40,000
60,000
80,000
100,000
120,000
140,000
160,000
1000 1500 2000 2500 3000
Solar field operating hours (h)
3500
a
500
a+b
1000 1500 2000 2500 3000
Solar field operating hours (h)
3500
Demand FWH #1
Demand FWH #1 + #2
Demand FWH #1 + #2 + #3
Site 2, case A: solar thermal energy distribution
500
Demand FWH #1
4000
4000
0
20,000
40,000
60,000
80,000
100,000
120,000
140,000
160,000
0
20,000
40,000
60,000
80,000
100,000
120,000
140,000
160,000
0
0
13.4 Solar thermal energy distribution in the selected locations.
0
0
a
a+b
Demand FWH #1 + #2
Demand FWH #1 + #2 + #3
Site 1, case A: solar thermal energy distribution
Thermal power (MW)
Thermal power (MW)
1000 1500 2000 2500 3000
Solar field operating hours (h)
Demand FWH #1
3500
500
a
b
c
1000 1500 2000 2500 3000
Solar field operating hours (h)
3500
Demand FWH #1
Demand FWH #1 + #2
Demand FWH #1 + #2 + #3
Site 2, case B: solar thermal energy distribution
500
a
b
c
Demand FWH #1 + #2
Demand FWH #1 + #2 + #3
Site 1, case B: solar thermal energy distribution
4000
4000
Annual fuel consumption
435,000
430,000
425,000
Site 1
Site 2
420,000
415,000
410,000
405,000
400,000
Base case
Plant efficiency
Efficiency (%)
42.50
42.00
41.50
41.00
40.50
40.00
Base case
431
440,000
Case A
Case B
Case A
CO2 emissions saved (ton/a)
Fuel consumption (ton/a)
Integrating a Fresnel solar boiler
Case B
CO2 emissions saved
50,000
40,000
30,000
20,000
10,000
0
Base case
Case A
Case B
13.5 Performance of the hybrid power plant coal energy conversion
efficiency, coal fuel consumption and CO2 emissions saved for Site 1
(DNI = 1,950 kWh/m2/year) and Site 2 (DNI = 2,300 kWh/m2/year).
The economic analysis of the solar add-on was based on a strategy that
integrates all revenues in an overall compensation tariff T:
T = Tfeed-in + FS + C CO2
where T is the overall compensation applicable to solar add-ons (c/kWh).
Tfeed-in is the feed-in tariff for solar thermal generation, which includes the
base electricity price (considering only the solar contribution) and the
incentives (c/kWh), FS is the component due to fuel savings derived from
the operation of the solar field (c/kWh) and CCO2 is the credit support due
to CO2 emission avoidance (c/kWh). Assumptions of the economic model
of the solar powered add-on are shown in Table 13.3.
Several assumptions were considered in order to select the values for the
base scenario.
• The feed-in tariff was selected to be lower than the current feed-in tariff
in Spain but higher than the generation cost.
• Fuel price is an average of the international coal price over 10 years.
• Although a fix CO2 emission avoidance credit has not been established,
the European Union Emissions Trading Scheme (EU ETS) has analysed
it and suggested a carbon credit of around 20 c/ton CO2 [7]. In order to
have a conservative calculation, 25% less was assumed.
© Woodhead Publishing Limited, 2012
432
Concentrating solar power technology
Table 13.3 Assumptions of the economic model of the solar powered
add-on
Financial assumptions
Total solar field cost
Equity ratio
Depreciation
Annual indexation of feed-in tariff
Inflation rate
Discount rate
c71,700,000
100%
25 years
2%
2%
6.5%
Base scenario
Feed-in tariff
Fuel price for savings
CO2 emission avoidance
0.16 c/kWh
60 c/ton
15 c/ton-CO2
Only the electricity generation contribution from solar steam generates
revenues according to Tfeed-in. The electricity produced conventionally is not
taken into account for revenues nor the fuel costs derived from it. The
economic calculation is focused on the cost of the equipment added to the
plant and its operation. All other operations and maintenance (O&M) costs
of the conventional plant are excluded from the analysis.
This model assumes that out of the total electricity generated by the
turbine, there is a fraction associated with the presence of the solar add-on.
The difference in fuel consumption attained when the solar field is operated
corresponds to a certain thermal power that is indirectly put into the system
by solar steam production. By means of the plant’s electric efficiency, this
is related to an equivalent ‘solar’ electric output, which is used to assess the
solar field’s effective generation costs. The results of the plant’s performance under the aforementioned scenarios are compared in Table 13.4.
The higher annual DNI site enables the add-ons to obtain a higher
thermal output, since more solar hours at the site let the system achieve a
higher solar share for the related electricity production. Both facts imply a
higher potential for fuel savings and thus for emission avoidance.
The use of a steam flow priority header (case A) is highly recommended
to allocate the thermal production of the solar field to achieve a higher
reduction in fuel use and thus a design with better economic performance.
With regard to site selection, generation costs are expected to be up to
0.02 c/kWh higher at the lower DNI site studied.
The inclusion of solar add-ons in the plant’s layout can be paid off in nine
years at the high DNI site under these assumptions, leading to a higher
internal rate of return (IRR). The reference framework assumes a feed-in
tariff of 0.16 c/kWh which allows solar field implementation within a reasonable IRR, electricity production cost and payback time ranges. The
© Woodhead Publishing Limited, 2012
© Woodhead Publishing Limited, 2012
a
c/kWh
c/kWh
c/kWh
c/kWh
0.013
0.022
0.160
0.195
0.013
0.022
0.160
0.195
c/kWh
c/kWh
c/kWh
c/kWh
c32,384,669
11,23%
9 years
0.152 c/kWh
c71.700,000
c44,949,859
12,91%
8 years
0.136 c/kWh
1.60%
1.42%
120
5% of total investment
88%
17,805 ton/a
15,771 ton/a
41,207 ton/a
36,500 ton/a
0.013
0.022
0.160
0.195
c/kWh
c/kWh
c/kWh
c/kWh
0.013
0.022
0.160
0.195
c/kWh
c/kWh
c/kWh
c/kWh
c21,916,192
9,78%
10 years
0.169 c/kWh
14,115 ton/a
32,668 ton/a
16,005 ton/a
37,043 ton/a
c33,530,952
11,38%
9 years
0.150 c/kWh
1.26%
Case B
1.43%
Case A
Case A
Case B
Site 2
Site 1
Assumed low heating value (LHV) of the coal =24,466 kJ/kg.
Financial results
Total investment
NPV
IRR
Discounted payback period
Electricity production cost
Revenues per kWh
→ CO2 crediting system
→ Fuel saving
→ Feed-in tariff
Total
Technical assumptions and results
Solar share of total generation
Collector modules
Major overhaul (13 years)
Boiler efficiency
Fuel savingsa
CO2 emission avoidance
Location
Table 13.4 Results of the economic model of the solar powered add-on by site
434
Concentrating solar power technology
Internal rate of return (%)
16
14
Site 1, case A
Site 1, case B
Site 2, case A
Site 2, case B
12
10
8
6
0.13
0.16
0.19
Feed-in tariff (€/kWh)
14
Internal rate of return (%)
Internal rate of return (%)
14
13
12
11
10
9
8
30
60
90
Price of fuel (€/ton)
13
12
11
10
9
8
10
15
20
Price of CO2 credits (€/ton-CO2)
13.6 Variation of the project IRR to equity.
advantages coming from fuel savings and CO2 credits represent a 0.035 c/
kWh additional support.
The sensitivity of the results to the revenue contributions is relevant to
knowing how profitable the implementation of the solar add-ons will be
under different scenarios (e.g., price variation of fuel and CO2 credits). The
results are shown in Figs 13.6 and 13.7. The range of the analysis was constrained to about ± 25% difference in feed-in tariff and to the latest trends
in coal prices. The price of carbon credits is likely to be adjusted according
to the country’s regulations.
Since the sale of solar electricity has the biggest share in the proposed
revenue system, the feed-in tariff is the driving force of the analysis. In the
range of about ±25%, the feed-in tariff can modify the IRR up to 2.5 percentage points. Likewise, it can impact the payback period by one to two
years, depending on the site rather than on the add-on’s configuration. The
economics of including solar add-ons in the plant layout are more sensitive
to the price of fuel saved than to the price of carbon credits. The changes
that the discounted payback time is subjected to by varying the price of fuel
or CO2 in some cases (see Fig. 13.7 (b) and (c)) are not perceived, since the
variations are smaller than one year.
© Woodhead Publishing Limited, 2012
Integrating a Fresnel solar boiler
435
Payback time (years)
13
12
11
Site 1, case A
Site 1, case B
Site 2, case A
Site 2, case B
10
9
8
7
6
0.13
0.16
0.19
Feed-in tariff (€/kWh)
12
Payback time (years)
Payback time (years)
12
11
10
9
8
7
30
60
90
Price of fuel (€/ton)
11
10
9
8
7
10
15
20
Price of CO2 credits (€/ton-CO2)
13.7 Variation of the discounted payback time.
13.4
Conclusions
The optimum scenario for the most economic introduction of solar add-ons
to existing power plants is, not surprisingly, for sites with the highest DNI.
Diverting the solar steam with priority to the HP feedwater heater (case
A) maximizes the use of the solar heat input, achieving a reduction of fuel
consumption that is equivalent to a rise in solar share of electricity generation of up to 0.2%.
The assumption to use a base load power plant for the case study leads
to a relative low solar share. For peak load plants, the solar add-ons will
have a higher impact on the overall performance of the plant, since their
capacity factor is lower and their time of operation is comparable to those
of the solar add-ons, avoiding the potential that solar production has to be
wasted when the power plant is out of operation.
A feed-in tariff as low as 0.16 c/kWh allows the implementation of solar
field add-ons in a payback time frame of 8–10 years in a high DNI site,
depending on the system configuration. In this context the IRR remains
close to 13%, even in case of price volatility of coal or carbon credits.
© Woodhead Publishing Limited, 2012
436
Concentrating solar power technology
The feed-in tariff of the assumed revenue scheme is up to 0.14 c/kWh
lower than, for instance, the current tariff for solar thermal in Spain (0.27 c/
kWh). This scheme includes the support provided by CO2 emission avoidance and the fuel savings (a characteristic advantage of the configuration
of the solar add-ons).
13.5
References
[1] Morin G., Lerchenmüller H., Mertins M., Ewert M., Fruth M., Bockamp S.,
Griestop T., Häberle A. (2004), ‘Plug-in strategy for market introduction of
Fresnel-collector’, SolarPaces Conference.
[2] Ugolini D., Zachary J., Park J. (2009), ‘Options for hybrid solar and conventional
fossil plants’, Bachtel Technology Journal, 2, 1, 1–11.
[3] Yang Y.P., Cui Y.H., Hou H.J., Guo X.Y., Yang Z.P., Wang N.L. (2008), ‘Research
on solar aided coal-fired power generation system and performance analysis’.
Science in China Series E: Technological Sciences, 51, 8, 1211–1221.
[4] Meteonorm, meteorological database developed by Meteotest, www.meteotest.
ch.
[5] Thermoflex, thermodynamic tool developed by Thermoflow, www.thermoflow.
com.
[6] Black & Veatch (1996), Power Plant Engineering, New York, Springer.
[7] Committee on Climate Change (2009),‘Meeting Carbon Budgets – the need for
a step change’, Progress report to Parliamentary Committee on Climate.
© Woodhead Publishing Limited, 2012
14
The long-term market potential of
concentrating solar power (CSP) systems
S. J. S M I T H, Pacific Northwest National Laboratory
and University of Maryland, USA
Abstract: This chapter will examine the conditions under which thermal
concentrating solar power (CSP) systems might play a larger role in the
global energy system during the twenty-first century. CSP technologies,
such as parabolic troughs or power towers, have a large advantage over
other solar technologies in that they offer the potential for firm power
delivery, mitigating intermittency issues. These systems require relatively
cloud-free conditions to operate, however, which limits their geographic
applicability.
Key words: concentrating solar power, electric generation.
14.1
Introduction
Concentrating solar power (CSP) systems are at a transition point in their
development. As discussed in Chapter 1, installed capacity in 2012 is close
to 2 GWe following several years of rapid growth. With average capacity
factors averaging around 40%, total annual energy generation is still only
of the order of one large coal-fired power station. Nonetheless, this is sufficient to establish CSP as a serious industry and continued growth is predicted. While the current contribution is relatively small, as the number of
CSP plants built increases, the expectation is that, following general historical trends in technology development, costs are likely to fall as the production of CSP plant components becomes more streamlined, operating
procedures are standardized, and plant designs improved.
As with most renewable technologies, the largest determinant of the cost
of electricity generated using CSP is the capital cost. The extent to which
these costs fall over time will, therefore, be a principal determinant of CSP
market share. Water availability, site characteristics, financing options, transmission capability, and the cost of alternative technologies will also all play
a role in determining CSP market potential. The role of these factors will
be discussed below. The general considerations below apply to all thermal
CSP technologies: trough, dish, tower and Fresnel; however, with parabolic
trough technologies being the most commercially mature and widely
deployed, some of the specific analysis is focused on trough systems.
437
Published by Woodhead Publishing Limited, 2012
438
14.1.1
Concentrating solar power technology
The role of concentrating solar power (CSP)
systems in the electric system
One key feature of thermal CSP plants is that they are generally constructed as hybrid units, whereby an auxiliary boiler or other heating unit
using natural gas,1 can be used to heat the working fluid and power the
generation system. With a relatively small additional capital expense, CSP
plants are, therefore, capable of producing predictable power. The auxiliary
system can be operated during a short cloudy period during the day or
during entire cloudy days when the system would otherwise not be operational. This provides an inherent advantage over many other renewable
energy technologies in that intermittency of the solar resource is mitigated
with little capital cost. Although fuel costs are incurred, which will be discussed later, CSP hybrid plants can be conceptually considered much like
conventional fossil electric generation technologies in terms of their dispatchability. The topics of energy storage and hybridization are covered in
detail in Chapters 11 and 12, respectively. Many commercial plants are now
also in operation incorporating large-scale thermal energy storage, which
allows solar operation into the evening.
CSP plants will operate within electricity markets that can be simplistically considered to consist of base, intermediate, and peak load segments.2
Base load plants are run nearly 24 hours a day, as exemplified by large coal
and nuclear plants. Intermediate plants are generally brought on line when
loads begin to increase during the day, with natural gas turbines (combined
cycle and conventional) often serving this load segment. Peaking plants
operate only during the highest demand periods, which in sunny regions is
generally during mid-day and evening in the summer. These power plants
are used only a relatively small fraction of the year. In the nomenclature
of the field, their capacity factor is low. This favors plants with low capital
costs even if efficiency is lower than in intermediate plants. Open cycle
natural gas or oil combustion turbines are generally used for the purpose
of supplying peaking power as these offer low capital cost and a short startup time.
CSP technologies are well suited to supply intermediate and base load
market segments. Most current CSP plants operate only during the day and
are best considered as intermediate load plants. Intermediate load comprises around 15–20% of total electricity demand in energy terms and this
1
Whilst natural gas has been used nearly exclusively for backup purposes to date, in principle
any fuel could be used, including biomass.
2
Markets in some jurisdictions (Eastern Australia, for example), have actually moved to the
concept of a continually varying cost of electricity based on supply and demand estimates for
short future time intervals.
Published by Woodhead Publishing Limited, 2012
The long-term market potential of CSP systems
439
defines the maximum market for CSP plants without large amounts of
thermal storage. As discussed further below, thermal storage is not necessarily essential to initial penetration in the intermediate electric power
segment, although storage does lower costs somewhat and expands the
maximum portion of intermediate load that can be economically served by
CSP technologies.
Base load generation is the largest market segment, comprising around
three-quarters of total electric generation in the United States, for example.
CSP plants with around 12 hours of thermal storage could provide base
load power, which offers the largest potential role for CSP technologies.
Note that base load CSP plants would still requir