KTH Energy and Environmental Technology
Exergoeconomic Analysis and Benchmark of a Solar Power Tower with Open Air
Receiver Technology
D I P L . - I N G . ( F H ) F E L I X U D O E R T L
Master of Science Thesis
KTH School of Industrial Engineering and Management Energy Technology EGI-2012-015MSC
Division of Applied Thermodynamics and Refrigeration SE-100 44 STOCKHOLM
Master of Science Thesis EGI 2012: 015MSC
Exergoeconomic Analysis and Benchmark of a Solar Power Tower with Open Air Receiver
Technology
Felix Udo Ertl
Approved 2012-03-20
Examiner
Prof. Dr. Björn Palm
Supervisor
Prof. Dr. Nabil Kassem
Commissioner Contact person
Thesis
This thesis has been written in conjunction with the German Aerospace Centre, Deutsches Zentrum Für Luft- und Raumfahrt e.V. (DLR), at the Institute for Technical Thermodynamics – Solar research and development, Cologne.
The thesis was examined by the Royal Institute of Technology, Kungliga Tekniska Högskolan (KTH), Energy and Environmental Technology, Stockholm.
Academic Supervisor: Prof. Dr. Nabil Kassem Academic Examiner: Prof. Dr. Björn Palm
Industrial Supervisors: Dipl.-Ing. Peter Schwarzbözl
Abstract
Concentrating solar power (CSP) provides a promising answer for the rising energy demand in emerging and developing countries as well as it can play a significant role for industrialized countries to provide renewable energy on demand. The highly efficient storage is the key component for CSP plants. Energy and exergoeconomic analysis are conducted in this work revealing potential improvements of a young and promising solar tower technology with open volumetric air receiver (OVR). Weaknesses and potential improvements are detected in the air cycle and receiver, in particular the absorber structure. The focus for improvements must further be on the cost reduction of the heliostat field rather than on improving its efficiency. A benchmark shows higher operation efficiencies for the parabolic trough system. This is because of the better performance of transferring heat with the oil cycle, even though this heat transfer cycle has obvious drawbacks, such as a large receiver area, lower temperatures and its inertia. In solar operation mode without additional burner the levelized electricity costs (LEC) are therefore 2.26 €cents lower for the trough system. The OVR tower, however, has clear advantages with an auxiliary burner due to the good adaptivity into the system.
The LECs are then reduced to 14.48 €cents/kWh, which is 6.5 €cents less as for the trough system with auxiliary burner. Significant improvements can be expected for the OVR tower when a gas turbine is deployed.
Keywords: Concentrated solar power, CSP, central receiver, open volumetric air receiver, power tower, thermal storage, solid bed storage, ceramic receiver, renewable energy, solar tower Jülich
Acknowledgment
I would like to express my sincerest appreciation to Peter Schwarzbözl, of the Institute for Technical Thermodynamics, Deutsches Zentrum für Luft- und Raumfahrt.
Without your sound advice, experienced insight and strong support throughout the duration of this thesis work the accomplishments would not have been possible.
I would like to send my gratitude to Sweden and the KTH that I was able to study such a valuable course there. Thanks to Prof. Dr. Nabil Kassem, my supervisor in the department of Energy Technology at KTH. I heartedly thank you for your crucial support for this thesis work as well as for the freedom which you offered me to manage and decide on the direction of the project.
I want to thank with all my heart my dear parents, Arthur and Renate Ertl, the best parents I could have had. You both have always believed in me and given a support throughout my life. Father, I will always carry you in my heart. I know you will help me in spirit when I am low.
AC Annual costs
CSP Concentrated solar power
DLR Deutsches Zentrum Für Luft- und Raumfahrt (German Aerospace Department) DNI Direct normal irradiation
ECOSTAR European Concentrated Solar Thermal Road-Mapping EES Engineering equation solver
EIA Energy Information Administration (US) ECO Economizer
HD Hochdruck (high pressure) HFLCAL Heliostat field layout calculator HRSG Heat recovery steam generator HTF Heat transfer fluid
HTX Heat exchanger IAM Incident angle modifier IDR Incident direct radiation IEA International Energy Agency LCC Life cycle cost
LEC Levelized electricity cost LHV Lower heating value
MD Mittlerer Druck (medium pressure) MENA Middle East and North Africa ND Niederdruck (low pressure)
NREL National Renewable Energy Laboratory NPV Net present value
O&M Operation and maintenance OVR Open volumetric receiver PSA Plataforma Solar de Almería PT Parabolic trough
PV Photovoltaic
RRM Revenue requirement method SEGS Solar Electric Generating System SM Solar multiple
TRR Total revenue requirement
VDI Verein Deutscher Ingenieure (Organization of German Engineers)
1. Introduction 2
1.1. Background . . . . 2
1.2. Motivation . . . . 3
1.3. Objective . . . . 4
2. State of the Art 5 2.1. Classification of CSP Technology . . . . 5
2.2. Overview of Existing Power Tower Technologies . . . . 6
3. Components of the Analyzed Plants 9 3.1. Parabolic Trough Power Plant . . . . 9
3.1.1. The Solar Field Including Receiver . . . . 10
3.1.2. Thermal Storage with Molten Salt . . . . 15
3.1.3. Power Block . . . . 16
3.2. The Solar Power Tower with Open Volumetric Air Receiver (OVR) Technology . . . 16
3.2.1. The Heliostat Field . . . . 17
3.2.2. The Open Volumetric Receiver . . . . 19
3.2.3. The Solid Matter Thermal Storage . . . . 20
3.2.4. The Power Block . . . . 21
4. Methodology 22 4.1. Exergy Analysis . . . . 22
4.1.1. Exergy . . . . 22
4.1.2. Exergy Balance . . . . 24
4.1.3. Exergy Destruction, Exergy Loss, and Exergetic Efficiency . . . . 25
4.2. Exergoeconomic Analysis . . . . 26
4.2.1. Aggregation Level . . . . 26
4.2.2. Thermoeconomic Variables for Component Evaluation . . . . 26
4.3. Thermoeconomic Evaluation . . . . 27
5. Modelling and Simulation 29 5.1. Operation Strategy . . . . 31
5.1.1. Location . . . . 31
5.1.2. Method of Operation . . . . 32
5.2. Power Plant Models . . . . 33
5.2.1. Parabolic Trough Model . . . . 34
5.2.2. Solar Tower . . . . 35
6. Simulation Results and Evaluation 42 6.1. Exergy Analysis . . . . 42
Contents
6.1.1. Steady State Calculations . . . . 42
6.1.2. Annual Performance Simulations . . . . 52
6.2. Exergoeconomic Analysis . . . . 54
6.2.1. Cost Calculations . . . . 54
6.2.2. Results and Evaluation of the Exergoeconomic Analysis . . . . 65
6.2.3. Optimizations and Modifications of the Solar Power Tower Plant . . . . 70
6.2.4. Annual Performance of the Optimized Systems . . . . 77
7. Discussion 79 7.1. Simulation in Ebsilon . . . . 79
7.2. Exergoeconomic Analysis . . . . 79
7.3. Cost Calculations . . . . 80
8. Conclusion 81 9. Outlook 82 Bibliography 83 A. Appendix - Simulation Characteristics 85 A.1. Control Diagram . . . . 86
A.2. Power Plant Specifications . . . . 87
A.3. Ebsilon Models . . . . 93
A.3.1. The Parabolic Trough Scheme of the Andasol Power Plant . . . . 93
A.3.2. The Open Volumetric Receiver Tower Scheme . . . . 94
A.4. Development of the Receiver and Solid Bed Storage for Ebsilon . . . . 95
A.4.1. The Receiver . . . . 95
A.4.2. The Solid Bed Storage . . . . 113
B. Appendix - Results 123 B.1. Steady State Simulation . . . . 123
B.2. Annual Performance Simulation . . . . 129
B.3. Exergoeconomic Results Ordered by Importance . . . . 131
1.1. Outlook of the annual electricity demand and generation in the EUMENA-region ac- cording to MED-CSP scenario. [Trieb, 2005] . . . . 2 1.2. The Athene Scenario, predicting cost drop with increased installed capacity and power
generation. [ECOSTAR] . . . . 3 1.3. The Solar Tower Power Plant in J¨ulich is the first of its kind. . . . 4 2.1. The main types of concentrating solar power: (a) Solar Fresnel and (b) Parabolic
Trough with a linear receiver concept; (c) Solar Power Tower and (d) Solar Dish with a central receiver concept [DLR] . . . . 5 3.1. Schematic illustration of a parabolic trough power plant with the energy flows over
the areas of the main components. . . . 9 3.2. The SEGS power plant located in Kramer Junction in the Mojave desert of California.
[ECOSTAR] . . . . 10 3.3. (a) The Eurotrough collector (b) Receiver Tube . . . . 10 3.4. The solar angles in relation to the surface of the earth. A is the azimuth angle (here
also determined asγ) and θzis the zenith angle. [Powerfromthesun] . . . . 11 3.5. Optical losses at the collector. Location: 30◦ north and 0◦declination/ Date: Septem-
ber 23th. [Sokrates, 2004] . . . . 13 3.6. Optical losses at the collector. Location: 10◦ north and 0◦declination/ Date: Septem-
ber 23th. [Sokrates, 2004] . . . . 13 3.7. Temperature characteristics of a solid matter storage. . . . 16 3.8. (a) A 100m2 glass-metal-heliostat with facets at the Plataforma Solar d`e Almeria.
[Sandia, 2010] (b) A 50m2SKI stretched-membrane heliostat.[Sandia, 2010] . . . . 17 3.9. Optical losses occurring in a heliostats’ solar field. . . . . 18 3.10. The receiver cubs are carried by a metal frame. The frame is cooled by the recycled
air flow (blue arrows) in order to withstand the high temperatures. (a) shows the frontal view of the receiver in J¨ulich and (b) the three parts, absorber cubs, metal frame structure and the hopper. Source: DLR . . . . 19 3.11. The volumetric effect causes higher temperatures behind the entrance of the absorber
channels. Source: DLR . . . . 20 3.12. Temperature characteristics of a solid matter storage. . . . 21 4.1. The system can interact with its environment to generate work. The exergy of the
system describes the maximum potential work that can be generated until the system and its environment are in equilibrium. [Bejan, 1996] . . . . 22 4.2. The balance of all flows over a control volume at steady state. Tb is the temperature
at which heat transfer occurs.[Bejan, 1996] . . . . 24 5.1. Data flow for the simulation of the open volumetric receiver tower plant . . . . 30
List of Figures
5.2. The simulation matrix gives an overview of the simulated power plant modifications. 31 5.3. Simplified operation mode of CSP plant with storage and auxiliary burner (solar hybrid). 32 5.4. The Ebsilon scheme for the parabolic trough power plant. . . . 33 5.5. The Ebsilon scheme for the open volumetric receiver tower. . . . 34 5.6. The influence of the field losses depending on the incident angle. . . . 36 5.7. The receiver cubs are carried by a steel frame. The frame is cooled by the recycled air
flow (blue arrows in figure b) in order to withstand the high temperatures. Figure (a) shows the Ebsilon model of the receiver split in functional parts. The numbers in the model show which part they represent in the scheme in figure (b). Figure (b) pictures the three parts, absorber cubs, metal frame structure and the hopper. . . . 37 5.8. The surface fitting tool diagram shows the accuracy of the polynomial equation in
comparison with the calculated results of the absorber efficiencies. The number356 represents the inlet temperature of the absorber in Kelvin. eta is the absorber effi- ciency,T aus the outlet temperature and Ic the incident concentrated solar energy on the absorber. . . . 39 5.9. The diagram contains values for the K-value and the upper temperature difference
that have been calculated with the EES receiver model for various mass flows. The polynomial equation calculated with Excel allows to adjust partial characteristics of the parts in the Ebsilon model. . . . 40 5.10. The air blower arrangement in the model is set up with a main blower and a storage
blower in parallel, where as the real system is set up with a receiver blower and a boiler blower in series. . . . 41 6.1. Energy and exergy efficiencies of the solar power tower in comparison: (a) Solar
Only, (b) With Auxiliary Burner. The abbreviations are explained in the table 6.1. . 42 6.2. Energy and exergy efficiencies of the parabolic trough in comparison: (a) Solar
Only, (b) With Auxiliary Burner. The abbreviations are explained in the table 6.1. . 43 6.3. Energy losses of the main parts of the solar power tower in comparison on a logarith-
mic scale: (a) Solar Only - Energy Losses and (b) With Auxiliary Burner - Energy Losses. . . . . 45 6.4. Exergy destructions of the main parts of the solar power tower in comparison on
a logarithmic scale: (a) Solar Only - Exergy Destructions and (b) With Auxiliary Burner - Exergy Destructions. . . . 46 6.5. Energy losses of the main parts of the parabolic trough in comparison on a logarith-
mic scale: (a) Solar Only - Energy Losses and (b) With Auxiliary Burner - Energy Losses. . . . . 47 6.6. Exergy destructions of the main parts of the parabolic trough in comparison on a
logarithmic scale: (a) Solar Only - Exergy Destructions and (b) With Auxiliary Burner - Exergy Destructions. . . . 48 6.7. A Sankey diagram of the exergy flow of the air cycle of the solar power tower in
operation mode SE1. Other Sankey diagrams are attached in Appendix B.1. . . . 50 6.8. A Sankey diagram of the exergy flow in the thermal oil cycle in operation mode SE1
under full load. . . . 51 6.9. One day of the solar tower in operation with low solar insulation and an empty storage,
simulated with an auxiliary burner and without. . . . 53
6.10. Comparison of the annual performance of the solar power tower with and without auxiliary burner. The diagrams show the annual exergy efficiency (equal to the energy
efficiency) and exergy destruction of the main parts. . . . 54
6.11. Cost distribution over the components of a Rankine cycle in a solar thermal power tower. 56 6.12. Cost distribution over the parts of the solar power tower (a) and the parabolic trough (c). 58 6.13. The gas burner configuration in the power plant scheme of the solar power tower (a) and the parabolic trough (b). The brown pipes represent a gas air mixture, the fuel pipe is pink and the thermal oil pipe is gray. . . . 60
6.14. Overview of the different optimized systems for investigation. . . . 71
6.15. A comparison of the annual exergy efficiencies of the solar tower and the parabolic trough with and without a duct burner. . . . 71
6.16. The diagram shows the behavior of the power block in the morning of a typical sum- mer. The storage is empty. Two configurations are displayed: The system supported by a duct burner and the system in solar only mode. . . . 72
6.18. Exergy efficiencies of operation modes with different HTF recovery factors with the support of a duct burner. . . . . 74
6.17. Overall exergy efficiencies of an annual performance simulation for the different con- figurations of the solar power tower in solar only operation. . . . 74
6.19. Annual exergy destruction including exergy losses in the main parts of the solar power tower for the different configurations. . . . . 75
6.20. The cost rate of the annual exergy destruction C of the different configurations and main parts in comparison. . . . 76
6.21. The cost importance C+Z (cost rate + cost of components) of the different configura- tions and main parts in comparison. . . . . 76
A.1. The logical control scheme for the csp plants simulated in this work. . . . 86
A.2. The Ebsilon scheme for the parabolic trough power plant. . . . 93
A.3. The Ebsilon scheme for the open volumetric receiver tower. . . . 94
A.4. The diagram contains values for the K-value and the upper temperature difference that have been calculated with the EES receiver model for various mass flows. The polynomial equation calculated with Excel allows to adjust partial characteristics of the parts in the Ebsilon model. . . . 112
A.5. The component representing the collecting hopper in the Ebsilon model is considering the heat losses that occur in the component. The condition under partial load is shown in the diagram. The polynomial equation is used in the Ebsilon model. . . . 112
A.6. The black curve represents the polynomial equation derived from the EES values. Additionally, this diagram includes a comparison with the polynom of adjustment suggested by Ebsilon. . . . 113
B.1. A Sankey diagram of the exergy flow in the heat transfer unit of the solar power tower in operation mode SE1. . . . 124
B.2. A Sankey diagram of the exergy flow in the heat transfer unit of the solar power tower in operation mode SNE3. . . . 124
B.3. A Sankey diagram of the exergy flow in the power block of the solar power tower in operation mode SE1 under full load. . . . 125 B.4. Solar Only - Energy and exergy efficiencies of the solar power tower in comparison. 126
List of Figures
B.5. With Auxiliary Burner - Energy and exergy efficiencies of the solar power tower in
comparison. . . . 126
B.6. Solar Only - Energy losses in comparison on a logarithmic scale. . . . . 127
B.7. With Auxiliary Burner - Energy losses in comparison on a logarithmic scale. . . . . 127
B.8. Solar Only - Exergy destruction in comparison on a logarithmic scale. . . . 128
B.9. With Auxiliary Burner - Exergy destruction in comparison on a logarithmic scale. . 128
B.10. Overall exergy efficiencies of an annual performance simulation for the different con- figurations of the solar power tower in comparison. . . . 130
B.11. Exergy destruction including exergy losses in the main parts of the solar power tower for the different configurations. . . . 130
B.12. The cost rate of exergy destruction C of the different configurations and main parts in comparison. . . . 131
B.13. The cost importance C+Z (cost rate + cost of components) of the different configura- tions and main parts in comparison. . . . . 131
5.1. Characteristic weather data and geographic data of Seville, which has been used for
simulations. Source: DLR . . . . 31
5.2. Overview of the operation conditions. . . . 32
5.3. Results of the solar field calculations with HFLCAL. . . . 35
5.4. The four parameters show the goodness of fit of the polynomial equation in figure 5.8. 39 6.1. Table of abbreviations for the simulations of all characteristic operations of the solar power tower. . . . 43
6.2. Energy and Exergy analysis of the receiver unit in operation mode SE1. . . . 49
6.3. Energy and Exergy analysis of the steam turbine and the condenser unit in operation mode SE1. . . . 51
6.4. Annual performance of the solar power tower and the parabolic trough with and with- out an auxiliary burner. . . . 52
6.5. Cost Calculation OVR Solar Power Tower . . . . 56
6.6. Cost Calculation Parabolic Trough . . . . 57
6.7. Operation Cost and Total Revenue Requirement of the Solar Power Tower . . . . 59
6.8. Operation Cost and Total Revenue Requirement of the Parabolic Trough . . . . 59
6.9. Detailed Cost Distribution of the OVR Solar Tower . . . . 63
6.10. LEC of the Power Tower . . . . 64
6.11. LEC of the Parabolic Trough . . . . 64
6.12. Solar Tower, Solar Only, @DP: recycling of the air stream: 60%, concentrated solar irradiation on the receiver: 500kW/m2, pressure of the water pipe after first econo- mizer: 1.5bar. . . . 66
6.13. Solar Tower, with Auxiliary Burner, @DP: recycling of the air stream: 60%, con- centrated solar irradiation on the receiver: 500kW/m2, pressure of the water pipe after first economizer: 1.5bar. . . . . 67
6.14. Parabolic Trough, Solar Only . . . . 69
6.15. Parabolic Trough, with Auxiliary Burner . . . . 70
A.1. Common specifications . . . . 87
A.2. Solar Parabolic Trough specifications . . . . 88
A.3. OVR Solar Power Tower . . . . 91
B.1. Table of abbreviations for the simulations of all characteristic operations of the solar power tower. . . . 123
B.2. Table of abbreviations for the annual performance simulations of the solar power tower. 129 B.3. Solar Only Basis, recycling of the air stream: 60%, concentrated solar irradiation on the receiver: 500kW/m2, pressure of the water pipe after first economizer: 1.5bar. . . 132
B.4. Solar only, recycling of the air stream: 80%, concentrated solar irradiation on the receiver: 500kW/m2, pressure of the water pipe after first economizer: 1.5bar. . . . . 135
List of Tables
B.5. Solar only, recycling of the air stream: 100%, concentrated solar irradiation on the receiver: 500kW/m2, pressure of the water pipe after first economizer: 1.5bar. . . . . 138
B.6. With auxiliary burner, recycling of the air stream: 60%, concentrated solar irradi- ation on the receiver: 500kW/m2, pressure of the water pipe after first economizer:
1.5bar. . . . . 141
B.7. With auxiliary burner, recycling of the air stream: 80%, concentrated solar irradiation on the receiver: 500kW/m2, pressure of the water pipe after first economizer: 1.5bar. 144
B.8. With auxiliary burner, recycling of the air stream: 100%, concentrated solar irradi- ation on the receiver: 500kW/m2, pressure of the water pipe after first economizer:
1.5bar. . . . . 147
B.9. Solar only, recycling of the air stream: 60%, concentrated solar irradiation on the receiver: 550kW/m2, pressure of the water pipe after first economizer: 1.5bar. . . . 150
B.10. Solar only, recycling of the air stream: 60%, concentrated solar irradiation on the receiver: 600kW/m2, pressure of the water pipe after first economizer: 1.5bar. . . . 153
B.11. Solar only, recycling of the air stream: 60%, concentrated solar irradiation on the receiver: 500kW/m2, pressure of the water pipe after first economizer: 3bar and 135◦C. . . . 156
B.12. Solar only, recycling of the air stream: 60%, concentrated solar irradiation on the receiver: 500kW/m2, pressure of the water pipe after first economizer: 3bar and 145◦C. . . . 159
B.13. Solar only, recycling of the air stream: 60%, concentrated solar irradiation on the receiver: 500kW/m2, pressure of the water pipe after first economizer: 3bar and 155◦C. . . . 162
1.1. Background
Concentrating Solar Power (CSP) technology is experiencing a revival the recent years driven by an increasing pressure on authorities due to the threat of global warming and the critical issue of energy security as well as advanced development achievements in the field of CSP technology.
The Desertec Initiative that gained momentum by the Club of Rome and the DLR is an example of the extensive potential of solar thermal power plants to meet a main share of the electricity demand in the near future. A joint-venture between 12 companies founded in 2009, namely the Desertec Indus- trial Initiative GmbH (DII GmbH), is already elaborating and negotiating the political and economical framework for an intercontinental grid tabbing the abundantly available renewable sources of energy in the EUMENA (EUropean, Middle East and North Africa) region. The Desertec Initiative aims to provide enough energy serving a substantial part of the energy demand of the MENA region and 15%
of Europe’s electricity demand from these energy sources by 2050 [DII]. The major source accounts to the solar irradiation in the MENA countries, which can be considered as the ”fuel” for CSP plants (see figure 1.1).
Figure 1.1.: Outlook of the annual electricity demand and generation in the EUMENA-region according to MED-CSP scenario. [Trieb, 2005]
In 1985, the Solar Electric Generating System I (SEGS I) the first commercial CSP plant, located in the Mojave desert in California, has been connected to the grid. It has a capacity of 13.8MWel and is ever since generating electricity. Eight further parabolic trough systems with a larger capacity followed the years after. All projects have been conducted by Luz International Limited. It was a
1.2. Motivation
great opportunity for the technology to prove its reliability and to become further developed.
Besides this early success story, which has been driven by the oil crises of the 70s, the cost of the electricity generated by any CSP technology is still the major remaining barrier for a widespread implementation. The electricity cost of solar thermal power plants is lying in the range of 15 - 20 centse/kWh today and is expected to drop to 5-7 centse/kWh according to the learning curve of the Athene Study in figure 1.2. [ECOSTAR]
Figure 1.2.: The Athene Scenario, predicting cost drop with increased installed capacity and power generation. [ECOSTAR]
Different ways can be followed in further technical development of CSP projects to induce a fast and effective cost reductions:
• Upscaling of existing technology can reduce Operation and Maintenance (O&M) costs as well as the component costs.
• New innovations relating technology and operational concepts.
• Understanding the complexity of the system and prioritizing the main cost drivers for a potential cost reduction.
In this work all three methods are applied. The potential of cost reduction by upscaling the solar tower with open air receiver relative to the Andasol parabolic trough power plant is investigated, new concepts of the power block are examined, and an exergoeconomic analysis enables a prioritized optimization of the solar tower.
1.2. Motivation
Solar thermal power plants use direct sunlight as an energy source with the aim of producing electricity with a conventional power plant process. Solar tower power plants are based on the concentration of direct solar radiation onto a solar receiver at the top of a tower using dual axis tracking mirrors (heliostats). In the receiver up to 1000-fold concentrated solar energy is absorbed and as heat at a very
high temperature level delivered to a heat transfer fluid (HTF). In the considered procedure ambient air as heat transfer fluid is heated in a ceramic volumetric receiver to about700◦C. The hot air is conducted to a heat recovery steam generator (HRSG) to drive a steam power process. Alternatively, the hot air is passed through a fixed-bed storage to store the heat.
Figure 1.3.: The Solar Tower Power Plant in J¨ulich is the first of its kind.
The development is currently in the demonstration plant stage. This plant is placed in J¨ulich in North Rhine-Westphalia in Germany with a capacity of 1.5MW electrical power generation. Kraftan- lagen M¨unchen, which develops this technology in collaboration with DLR, is planning to construct the first exclusively commercial solar tower with open air receiver in Algeria. The plant will have an upscaled capacity of presumably 15MW electrical output.
1.3. Objective
The performance and properties of the entire system as well as individual parts in an upscaled plant are underlying various changes compared to the demonstration plant. In particular these parts and subsystems that have greater impact on the overall performance should lie in the focus of develop- ment. Further, an optimization of the operational conditions of the power plant is stringent.
The objective of this thesis is to identify and analyze optimizing potentials of the upscaled system and subsystems of a solar thermal power tower. Simulation models are developed to analyze and optimize the overall system of the air receiver and the steam cycle.
Thereby, the following issues shall be addressed:
• A technical inside of the two concepts: the parabolic trough and the solar tower with open volumetric receiver
• The maximum possible system efficiency for a 50MW power plant
• Optimal operation in the solar-only or hybrid operation
• Comparison with the parabolic trough system
• Investigation and evaluation of various improvements
2. State of the Art
2.1. Classification of CSP Technology
In solar thermal power technology, heat gained from concentrated solar irradiation is supplied to a thermodynamic cycle process to generate electricity. There are four different types of solar thermal power technology existing and under development. The solar fresnel, parabolic trough, solar power tower, and the solar dish concepts are state of the art technology and already widely deployed (see figure 2.1). All of them are also determined as concentrating solar power, because solar irradiation is bundled by mirrors on a linear or punctual receiver.
(a) (b) (c) (d)
Figure 2.1.: The main types of concentrating solar power: (a) Solar Fresnel and (b) Parabolic Trough with a linear receiver concept; (c) Solar Power Tower and (d) Solar Dish with a central receiver concept [DLR]
The solar tower has a concentration ratio of up to 1000 and the solar dish up to 2000, whereas the fresnel and the parabolic trough system have a rather low concentration ratio of 40 and 90 respectively.
Higher concentration results in higher temperatures being absorbed at the receiver (see Eq. 2.1).
Tabs= TSun 4
r C
Cmax
(2.1)
Tabs temperature at absorber [K]
TSun = 5760K, temperature of the sun surface [K]
C concentration ratio [-]
Cmax = 46200, maximal concentration [-]
Central receiver systems benefit from a higher concentrating factor with higher temperatures in- duced into the system. Power towers reach a temperature of 1100◦C max. From a thermodynamic
point of view, higher temperatures have a significant advantage due to a greater possible overall effi- ciency according to the Carnot efficiency [Pitz-Paal, 2008]:
ηc= 1 − Tmin
Tmax (2.2)
However a high efficiency does not stringently mean lower electricity production costs, which are the overall goal in the improvement and adaptivity of these technologies.
2.2. Overview of Existing Power Tower Technologies
As explained above central receiver systems have the highest potential for competitive electricity gen- eration costs. However, none of the existing technologies has been proven as the leading technology concept yet. An overview of the existing tower concepts is helping project developers to get an un- derstanding about the wide field and the potential.
The concepts can be divided into 7 categories that are defined by the difference of at least one of its main components, which are the concentrators, the receiver and the thermodynamic cycle. The following pictures give a short introduction into the different technologies:
Pressurized Gas Receiver: The Solar Brayton Tower concept integrates a pressurized air re- ceiver and a pressurized solid media storage in a Brayton cycle. The exhaust of the gas turbine can be used for cogeneration. [Giuliano, 2010]
Pressurized Gas Receiver: The Solar Hybrid Combined Cycle (SHCC) combines several solar Brayton cycle towers with one Rankine cycle. [Giuliano, 2010]
Pressurized Gas Receiver: The CO2 Gas Tower comprises a cavity receiver with metal tubes using CO2 as HTF and a Rankine cycle.
[Giuliano, 2010]
Open Volumetric Receiver: The Open Volu- metric Receiver Tower is determined by its receiver that heats up ambient air, which is sucked through the ceramic receiver structure.
The thermodynamic process is a Rankine cy- cle.
2.2. Overview of Existing Power Tower Technologies
Direct Steam Receiver: The Direct Steam Single Tower generates saturated or super- heated steam directly in a tube receiver.
Direct Steam Receiver: The Direct Steam Multiple Tower concept has multiple medium size towers that are connected together to drive one steam turbine.
Direct Steam Receiver: The Direct Steam eS- olar concept is designed similar the multiple tower concept, but its unique selling points are mass produced components. Thus, mirrors and towers are smaller and the concentrator field is an uniformly distributed array.
Molten Salt Receiver: The Molten Salt Tower is one of the earliest concepts with a circular molten salt receiver. The molten salt is both storage medium and HTF. [Giuliano, 2010]
Beam Down Concentrator: The Tokyo-Tech Beam-Down Tower consists of a secondary concentrator on top of the tower and a molten salt receiver on ground level.
Beam Down Concentrator: The Multi- Purpose Solar Tower is combining the beam down with the direct steam concept. The molten salt receiver on ground level functions as a storage.
Solid Media Receiver: The Particle Tower;
Solid media particles are exposed to con- centrated sunlight in the receiver. The par- ticles heat air for a gas turbine in a di- rect heat exchanger and act as storage media.
[Giuliano, 2010]
Solid Media Receiver: The Graphite Tower;
A graphite block is the receiver and the thermal storage. Besides the induced solar thermal en- ergy, surplus electricity can be stored in form of thermal energy.
Emerging Technology: Solar Tiles for CSP in Cities; This concept allows to make use of ex- isting building surfaces to concentrate sunlight on one receiver. The solar tiles are integrated in the building and self-powered.
3. Components of the Analyzed Plants
3.1. Parabolic Trough Power Plant
Parabolic trough power plants consist mainly out of four parts: The solar collector field, the auxiliary burner, the thermal storage and the power block (see figure 3.1). The parabolic mirrors bundle the solar irradiation up to 90 fold onto the linear receiver, in which the heat transfer fluid, a thermal oil, is heated. Heat exchangers transfer the heat to the steam cycle of the power block. A detailed description of its function is given in the following chapters.
Figure 3.1.: Schematic illustration of a parabolic trough power plant with the energy flows over the areas of the main components.
This system relies already on a relatively long history of power generation with the first commercial power plants being built from 1985 to 1991, namely the SEGS power plants I to IX in the US (see figure 3.2). New power plants of this kind, like Andasol 1,2 and 3 in Spain, are still based on the same system, because the technology has proven to be commercially ready. Therefore, it serves as a proper basis for comparison in an exergoeconomic analysis with the solar tower technology.
Figure 3.2.: The SEGS power plant located in Kramer Junction in the Mojave desert of California.
[ECOSTAR]
3.1.1. The Solar Field Including Receiver
Direct solar radiation is focused on a linear absorber tube located in the focal line of a parabolic mirror reflector. The assembly is determined as solar collector. The solar field is arranged in parallel rows of several solar collectors. Two rows are connected as one loop, through which synthetic thermal oil is pumped to deliver heat to the steam cycle. The one axis tracking system follows the sun from sunrise to sunset. A 50MWel power plant using Eurotrough collectors with a dimension of 6m aperture width and 150m length has an effective field area of ca. 554 000m2.
(a) (b)
Figure 3.3.: (a) The Eurotrough collector (b) Receiver Tube
Field Losses
Geometrical losses, optical losses and thermal losses occur at the collector. These losses define the field efficiency. Losses due to heating up of the collectors are not considered here. (see Eq.3.1).
[Sokrates, 2004]
ηf ield = ηgeo· ηopt· ηtherm (3.1)
3.1. Parabolic Trough Power Plant
3.1.1.1. Geometrical Efficiencyηgeo
The geometrical efficiency is calculated with the quotient of the Incident Direct Radiation (IDR) on the mirror of the collectors and the actual Direct Normal Irradiation (DNI) (see eq. 3.2).
ηgeo= IDR
DN I (3.2)
The DNI is the normal solar irradiance in Watt per square meter on the surface of the earth. The IDR is the effective useful irradiation per square meter collector surface. There are several geometric parameters that influence the effective irradiation (eq. 3.3):
IDR = ζCos· ζIAM· ζAV · ζEV · DN I (3.3) Losses included in the IDR are depending on the solar angles. The figure 3.4 shows the relation of the angles for the following interpretations of the losses.
Figure 3.4.: The solar angles in relation to the surface of the earth. A is the azimuth angle (here also determined asγ) andθzis the zenith angle. [Powerfromthesun]
Cosine LossesζCos:
As the solar collector tracking axis is always oriented to the south-north axis, the incident radiation is not always perpendicular to the solar field. Just the part that is perpendicular can be reflected to the absorber. The losses are described by the cosine effect:
ζCos= cosθ (3.4)
Whereas, the angle of incidenceθ is given by the equation 3.5 according to [Duffie, 1991]:
cosθ = cosβ · cosθZ+ sinβ · sinθZ· cos(γC− γS) (3.5)
β angle of inclination between the collector area and the horizontal (=track angle) [◦]
θZ zenith angle of the sun [◦]
γS azimuth direction angle of the sun, [◦] γS= 0◦for south, east positive, west negative
γC azimuth angle of the collector [◦]
For collectors with an orientation from north to south it is:
γC = 90◦ forγS > 0◦ γC = 270◦ forγS < 0◦
The influence of the cosine effect becomes larger with increasing distance from the equator and it is at its maximum at solar noon, when the sun is at azimuthγ = 180◦(see figure 3.5).
Incident Angle Modifier LossesζIAM:
The IAM incorporates the fact that the solar radiations are not exactly parallel due to the finite dis- tance between earth and sun. So the reflection of the sun onto the receiver is not exactly circular and its ellipsoidal form variates slightly with a changing incident angleθ. Thus, a part of the reflected radiation is lost, because the size of the absorber is adjusted to the optimal reflection. An ideal cir- cular reflection occurs forθ = 0. According to [Marco, 1995] the losses related with the IAM are approximately calculated as follows:
ζIAM = cosθ · (1 + sin3θ) (3.6)
Shading LossesζAV:
Parallel collectors shade each other if the sun is near the horizon. This effect depends on the track angle (=transversal angle) of the parabolic collectors (see eq. 3.7) [Sokrates, 2004].
ζAV = 1 − 1 − R
A cosβ +tanαcosγsinβ
!
· 1 − A · sinβtanαsinγR A cosβ +tanαcosγsinβ L
!
(3.7)
R Distance of collector rows. [m]
A Height of collectors [m]
L Length of collector rows [m]
γ relative azimuth of collectors (|γS− γC|) [◦]
Optical End-lossesζEV:
End-losses are considered as the fraction of radiation at the end of a collector that doesn’t hit the receiver tube. These losses occur at an incident angleθ greater or less than 0◦, when the radiation is reflected with an angle to the vertical axis of the collectors. Since collectors are arranged in rows, the end-losses are diminished by the corresponding neighbor collector. End-losses are determined as follows [Sokrates, 2004]:
ζEV = 1 −|f · tanθsin|γS− γC||
L (3.8)
3.1. Parabolic Trough Power Plant
All optical losses occur related to the position of the sun varying during day and year and to the distance of the location from the equator. The diagram 3.5 and the diagram 3.6 show the variation of optical losses at different locations.
Figure 3.5.: Optical losses at the collector. Location: 30◦north and 0◦declination/ Date: September 23th. [Sokrates, 2004]
Figure 3.6.: Optical losses at the collector. Location: 10◦north and 0◦declination/ Date: September 23th. [Sokrates, 2004]
3.1.1.2. Optical Losses At Mirror And Receiver Tubeηopt
Just a fraction of the incident direct radiation is finally delivering heat to the absorber tube. Optical losses of the components of the collector are the source of further losses (see eq. 3.9) [Sokrates, 2004].
˙qA= IDR · δ · ρ · τM2 · γ · τC · α = IDR · ηopt (3.9)
δ Pollution factor; Pollution of the mirrors reduces the reflect-ability in average [-]
about 2%:δ = 0.98
ρ Reflection factor of the mirror; A full reflection is physically not possible:ρ = 0.93 [-]
τM Transmission factor of the mirror; The mirror glass covering the reflecting layer absorbs [-]
a small portion of the incident and reflected radiation:τM = 0.99
γ Quality factor; It incorporates surface failures, due to manufacturing inaccuracy, [-]
and misalignment of the mirror axis and absorber axis:γ = 0.90
τC Transmission factor of the cladding tube; Absorption occurs at the glass of the cladding [-]
tube of the absorber:τC = 0.95
α Absorption factor of the absorber tube; Radiation on the absorber tube is reflected and [-]
lost. Surface coating, such as black chrome dioxide or a high selective layer of a metal oxide ceramic basis improves the absorption to approximatelyα = 0.95.
3.1.1.3. Thermal Lossesηtherm
Additionally to transmission and absorption losses at the receiver tube, thermal losses through con- vection and subsequent thermal radiation from the absorber tube are significantly influencing the field performance.
Losses Through Thermal Radiation ˙qEm
According to [Kleemann, 1993] a simplifaction (see eq. 3.10) based on the thermal emission law of a black body can be used to identify the thermal radiation losses of the receiver tube. Manufacturer of receiver tubes have developed a selective metal-ceramic coating (Cermet) that facilitates a relative good absorption at a low emission rate.
˙qEm= π · σ · ǫ
C T4− TC4
(3.10)
σ Boltzmann constant:σ = 5.67 · 10−8 W
m2K4
ǫ Emission coefficient of the absorber tube; This factor provides the fraction of [-]
the heat emitting related to a black body at same temperature.ǫ = 0.16
C Concentration factor of the collector; For Eurotrough collectors is C=90 [-]
T Mean temperature of the heat transfer fluid [K]
TC Temperature of the the cladding tube: simplified asTC = Tambient+ 37K [K]
A factor, that has significant influence on the emission and that is not directly related with the re- ceiver tube, is the collector’s concentration factor of the isolation C. The new collector ”Eurotrough”
has reduced the influence of the radiation losses by providing a higher concentration rate of 90 com- pared to the ”LS2” collector with a concentration rate of 70. Collectors under development aim for even higher concentration rates, such as the Helio-Trough and the Ultimate-Trough.
Convection Losses ˙qConv
Vacuum is implemented in the annular space between absorber tube and cladding tube so that convec- tion is by approximation avoided. However, convection is occurring at the ends of the receiver, where
3.1. Parabolic Trough Power Plant
the cladding tube is mounted with an aluminum connection on the absorber tube (see figure 3.3):
˙qConv = π · U
C (T − TAmb) (3.11)
U heat loss coefficient; It is an empirically identified coefficient that takes all linear W
m2K
losses into account in particular convection at the receiver tube and the insulated
pipes: U=2W/m2K
Tamb Ambient temperature [K]
Forced convection as well as defocusing of collectors due to wind is not considered in this work.
3.1.2. Thermal Storage with Molten Salt
The ability of storing thermal energy in an efficient way, is seen as the most important asset of CSP- plants to all other energy technologies of solar sources1. The thermal storage in a solar power plant allows operation after sun set. It is a puffer during fluctuating sun irradiation caused by passing clouds, for instance. Furthermore, it offers the advantage of highly dispatchable energy that enables to generate electricity whenever the demand and thus the tariff of electricity is high. These advantages add an important economic factor to the CSP plant that should be utilized to an optimal extent.
The three different physical principles of thermal storage:
• Sensible heat storage (SHS): direct and indirect
• Latent heat storage (LHS) using phase change materials (PCM)
• Thermo-chemical heat storage (TCHS) using the principle of chemical reactions
The two types of storages applied in the power plant designs of this work are both storing indirect sensible heat. Thus, the sensible heat storage is depicted in the following:
A direct system is storing the heat transfer fluid, which is receiving solar heat in the receiver and generating steam in the boiler, directly in a tank. This is implemented in the ”solar tres” demonstra- tion plant in Andalusia in Spain, for instance, where a storage of 6,250 tons of molten nitrate salt is delivering heat up to 15 hours. The nitrate salt is stored directly without any heat transfer occurring.
[Sener, 2007]
An indirect system is defined with heat transfer occurring between the fluid or matter that stores thermal energy and the heat transfer fluid that collects thermal energy in the receiver. The indirect storage of the parabolic trough plant with thermo oil as heat transfer fluid is a molten nitrate salt storage. The molten salt is pumped through a heat exchanger between a cold and a hot tank. The storage has to be operated in a temperature range between 290 and 565◦C. If the temperature is reaching a critical low temperature, the salt must be heated to prevent it from freezing. Molten salt storages have proved to be reliable and high efficient, but they are still expensive and difficult to handle because of the molten salt characteristics, which are in particular the corrosiveness, the environmental hazard, and the freezing below 260◦C.
1Wind and wave energy is included in this denotation, because solar irradiation on earth causes temperature gradients and thus wind.
3.1.3. Power Block
The power block of the parabolic trough plant model generated for this work is similar to the concept of the Andasol power plants in Spain. It consists of a single-pressure HTF/ steam generation system, a Rankine steam turbine/ generator cycle, a dry cooling system and a six-fold feed-water preheater system. Different to the Andasol design is the dry cooling system. It causes higher investment costs and a lower overall cycle efficiency, but it allows the power plant to be operated in regions of scarce water, where solar irradiation is high.
A great asset of this power block is the use of thermal oil delivering the heat to the steam cycle. The thermal oil has a high heat transfer capacity and thus implies efficient and compact heat exchanger with low pinch point temperatures. However, due to the temperature limit of the HTF (395◦C), the superheated steam entering the steam turbine is of low temperature.
3.2. The Solar Power Tower with Open Volumetric Air Receiver (OVR) Technology
The plant is determined by its receiver, which contains cubs with an open volumetric ceramic struc- ture. Similar to the parabolic trough technology, the plant comprises concentrators, a receiver, a heat transfer fluid cycle with an auxiliary burner, a storage and a Rankine cycle power block 3.7.
Figure 3.7.: Temperature characteristics of a solid matter storage.
The innovation of this technology is the receiver that allows the use of ambient air as heat transfer fluid. When the air is sucked trough the receiver structure high temperatures of up to 800◦C can be achieved that allow better performance of the Rankine steam cycle. The recycled air (150◦C) is
3.2. The Solar Power Tower with Open Volumetric Air Receiver (OVR) Technology
leaving the cycle through the receiver to cool down the receiver structure. About 60% of the air leaving the receiver is recovered, because it gets sucked back in. A ceramic packed bed storage can be charged by hot air flowing in one direction and discharged in the opposite direction. The storage and the additional duct burner are compensating fluctuation of the solar irradiation.
After the successful research projects, TSA starting 1993 and Phoebus at the Plataforma Solar d`e Almeria (PSA), to develop and test different open volumetric receiver concepts, Kraftanlagen M¨unchen GmbH accomplished to build this concept in J¨ulich in 2008 in collaboration with the DLR.
The general feedback of the performance of the power plant has been positive.
3.2.1. The Heliostat Field
The mirrors in a solar field for central receiver plant concepts are named Heliostats. It originates from the Greek word helios for sun and stat for fixed basis. Heliostats are tracking the sun biaxial so that the reflected sunlight is always fixed to one place independent of the position of the sun. The focus of all heliostats is the receiver mounted on top of the tower.
(a) (b)
Figure 3.8.: (a) A 100m2 glass-metal-heliostat with facets at the Plataforma Solar d `e Almeria.
[Sandia, 2010] (b) A 50m2SKI stretched-membrane heliostat.[Sandia, 2010]
There is two different types of heliostats, the glass-metal-heliostats with facets of mirrors and the metal-membrane-heliostats (see figure 3.8). A common size ranges from 1 to 150m2. In contrary to the parabolic mirrors, each heliostat is tracked by its own actuators controlled by a central processing unit.
3.2.1.1. The Field Efficiency
The field efficiency is the factor that determines the energy of solar radiation that hits the receiver of the tower. This incident energy on the receiver is defined as follows [Pitz-Paal, 2008]:
Q˙rec = ˙Qsolar· ηf ield (3.12a) with : ˙Qsolar = DN I · AH· nH (3.12b) where AH is the area of one heliostat and nH is the number of all heliostats in a field. The field efficiency again contains different losses of the field:
