A Techno-Economic Feasibility Study On The Use of

A TECHNO-ECONOMIC FEASIBILITY STUDY ON THE
USE OF DISTRIBUTED CONCENTRATING SOLAR

POWER GENERATION IN JOHANNESBURG
Christiaan César Bode
A project report submitted to the Faculty of Engineering, University of

the Witwatersrand, Johannesburg, in partial fulfilment of the
requirements for the degree of Master of Science in Engineering.
Johannesburg, 2009
The financial assistance of the South African National Energy Research Institute
towards this research is hereby acknowledged. Opinions expressed and conclusions
arrived at, are those of the author and are not necessarily to be attributed to SANERI.
i
DECLARATION
I declare that this project report is my own unaided work. It is being submitted for
the Degree of Master of Science in Engineering in the University of the
Witwatersrand, Johannesburg. It has not been submitted before for any degree or
examination in any other University.
_____________________________________________________
(Signature of Candidate)
On this _________day of _________________________________ 2009
ii
ACKNOWLEDGEMENTS
I would like to express my sincere gratitude to all those who provided support and
assistance during the compilation of this report.
My supervisor, Professor T.J. Sheer, whose guidance is greatly appreciated.
Professor W. Cronje, of the School of Electrical and Information Engineering at the

University who initiated the idea and stimulated my research.
Mr E. Brink, of the Electrical Engineering department at the University, for his

assistance with Matlab code.
My parents, Lillias and Peter and my brother Sebastian, who endured my stress and
provided much needed support during the research.
Finally to Mr Thomas Roos, DPSS, CSIR, whose passion for Concentrating Solar
Power has been inspiring. With his optimism and zeal, these technologies will
become a reality in South Africa and his input and assistance have been
immeasurable.
iii
ABSTRACT
This study provides an evaluation of Concentrating Solar Power (CSP) technologies

and investigates the feasibility of distributed power generation in urban areas of
Johannesburg. The University of the Witwatersrand (Wits) is used as a case study
with energy security and climate change mitigation being the main motivators.
The objective of the study was to investigate the potential of CSP integration in
urban areas, specifically investigating Johannesburg’s solar resource. This is done by
assessing the performance and financial characteristics of a variety of technologies
in order to identify certain systems that may have the potential for deployment.
To aid the comparison of the technologies, CSP performance and cost data which
were taken from multiple sources, were adjusted giving it local, present day
assumptions. A technology screening process resulted in the conception of twelve
alternative design configurations, each with a reference capacity of 120 kW(e).
Hourly energy modelling was undertaken for Wits University’s West Campus for
each of the twelve alternatives. Three configurations were further investigated and
are listed below; each with a design capacity of 480 kW(e).
1. Compound Linear Fresnel Receiver (CLFR) field with an Organic Rankine

Cycle (ORC).
2. Compound Linear Fresnel Receiver field with an Organic Rankine Cycle that
integrates storage for timed dispatch.
3. Compound Linear Fresnel Receiver field with an Organic Rankine Cycle that
integrates hybridisation with natural gas.
Levelised electricity costs (LEC) of the systems were used as the basis for financial
comparison. Real LECs, for the three configurations above, range between
R4.31/kWh(e) (CLFR, ORC) and R3.18/kWh(e) (CLFR, ORC with hybridisation).
iv
With the energy modelling of the hourly direct normal irradiation (DNI) input into
the CSP systems, Wits University’s West Campus Electricity bill was recalculated.
The addition of the solar energy input resulted in certain savings and a new LEC that
is Wits-specific. These LECs ranged between R3.98/kWh(e) (CLFR, ORC) and
R2.77/kWh(e) (CLFR, ORC with hybridisation). A third LEC was calculated that
integrates a CSP feed-in tariff (REFIT) of R2.05/kWh. At the time of writing, a CSP
REFIT of R2.10/kWh was released which favours the analysis.
The analysis of the 480 kW(e) systems resulted in total plant areas of between
10350 m2 (CLFR, ORC,) and 15270 m2 (CLFR, ORC, with storage). With plant
modulation, these plants can be placed on vacant land, above parking lots or on top
of buildings which would also provide shading.
The values obtained for the average yearly insolation was 1781 kWh/m2 based on
TMY2 data. Johannesburg has a very intermittent source of DNI solar energy. The
summer months in Johannesburg yield a higher peak DNI, whereas the winter
months provide a more consistent average. This is due to the high amount of cloud
cover experienced in summer. With this insolation, CSP electric generation is
possible however, compared to the other locations, it is not ideal. Also, because of its
intermittency is has been advised that certain applications such as HVAC and
process heat and steam requirements be pursued.
From the results, it can be concluded that power production costs through small
scale CSP systems are still higher than with conventional fossil fuel options,
however several options that may favour implementation were recognised. Through
the analysis it was found that if the CSP generated electricity is valued at the market
price ( CSP REFIT), the payback time of such systems can be decreased from 73 to
12 years (CLFR, ORC with storage). Further, due to the scale of the plants analysed,
the exploitation of high efficiencies and economies-of-scale of plants with power
levels above 50 MW(e), is not possible. With the introduction of these technologies
v
at lower power levels, cost savings through the incorporation of other design
options (such as waste heat utilisation) should be pursued.
It was recognised that South Africa in general has one of the greatest solar resources
in the world and should therefore be technology leaders and pioneers in CSP
technology. With greater emphasis being placed on the need for renewable energy
systems, it is imperative that South Africa develops its skills and a knowledge base
that will work at making the implementation of renewable energy, and in particular
CSP generation, a reality. Technologies identified that should be pursued for
distributed generation include Linear Fresnel collectors that are easy to
manufacture and don’t involve complicated receiver systems. There is also scope for
developing thermal storage technologies in order to make generation more reliable.
vi
TABLE OF CONTENTS
DECLARATION ii
ACKNOWLEDGEMENTS iii
ABSTRACT iv
TABLE OF CONTENTS vii
LIST OF FIGURES xi
LIST OF TABLES xiii
LIST OF SYMBOLS AND ABBREVIATIONS xv
1 INTRODUCTION 1
1.1 Background 1
1.2 Motivation 2
1.3 Objectives 3
2 LITERATURE REVIEW 5
2.1 Utility and Distributed Generation 5

2.2 CSP Technology: Basic concepts 6
2.2.1 Introduction 6
2.2.2 Solar Energy Resource 7
2.3 Collector Types 9
2.3.1 Parabolic Trough Collector System 9
2.3.2 Compound Linear Fresnel Reflector (CLFR) 11
2.3.3 Central Receiver Technologies 13
2.3.4 Dish-Stirling Systems 14
2.3.5 Solar Chimney Technology 15
2.4 Variations in Design and Common Technologies 16
2.4.1 Storage 16
2.4.2 Hybrids 18
2.4.3 Integrated Solar Combined Cycle System (ISCCS) 19
2.4.4 Direct Steam Generation (DSG) 20
2.4.5 Organic Rankine Cycles (ORC) 21
2.5 Data Sources 22
2.5.1 Solar Electric Generation – A Comparative Overview (1997) 23
vii
2.5.2 Eskom CSP Pre-feasibility Study (2001) 25
2.5.3 Modular Trough Power Plants (MTPP) (2001) 26
2.5.4 Solarmundo line focussing Fresnel collector (2003) 26
2.5.5 Assessment of Parabolic Trough and Power Tower Technology (2003)
27
2.5.6 European Concentrated Solar Thermal Road-Mapping Report (2005)
30
2.5.7 The Present and Future Use of Solar Thermal Energy (2005) 33
2.5.8 California Studies for NREL (2005, 2006) 33
2.5.9 Assessment of the World Bank Group/GEF Strategy (2006) 35
3 METHODOLOGY 37
3.1 Outline 37
3.2 Data Synthesis 39
3.2.1 CSP Technical Performance 39
3.2.2 Financial Calculations 42
3.2.3 Model Development and Verification 44
3.2.4 Model Adjustment 45
3.3 Technology Screening 49
3.4 Application 51
3.4.1 CSP Plant Performance 52
3.4.2 Energy Modelling 56
4 TECHNOLOGY SCREENING 57
4.1 Economic Comparison of Existing Technologies 57

4.2 Candidate Technologies 60
4.3 Functional Criteria 61
4.4 Numerical Analysis 64
4.5 Perspective Model 66
4.6 Alternative Technology Evaluation 67
4.7 Chosen Alternatives and Discussion 69
5 APPLICATION 73
5.1 Wits Electricity Profiles 73

5.2 Site 74
5.3 DNI Data 75
5.4 Design Configurations 76
5.4.1 Output Capacity 77
viii
5.4.2 Reference Plant 77
5.4.3 Storage 78
5.4.4 Hybridisation 78
5.5 CSP Plant Performance 79
5.5.1 Parabolic Trough 79
5.5.2 CLFR 81
5.5.3 Thermal Energy Flow 82
5.6 Income and Expenses 85
5.6.1 Specific Costs 85
5.6.2 Levelised Electricity Cost 86
5.6.3 Income Sources 89
6 ENERGY MODELLING 94
6.1 DNI Synthesis 94

6.2 Design Analysis and Thermal Modelling 94
6.3 System Integration and Bill Calculation 96
7 RESULTS 99
7.1 Initial Design Results 99

7.2 Initial Financial Results 100
7.3 Matlab Modelling 103
8 DISCUSSION 113
8.1 Suitability of Design Approach 113

8.2 Results 117
8.3 Recommendations for Implementation 120
9 CONCLUSIONS AND RECOMMENDATIONS 126
9.1 Conclusions 126

9.2 Recommendations 129
REFERENCES 132
APPENDIX A STORAGE CONCEPTS 140
APPENDIX B TECHNOLOGIES USED IN THE COMPARISON 142
APPENDIX C SOLAR RADIATION DATA 150
APPENDIX D DATA VERIFICATION 153
ix
APPENDIX E JOHANNESBURG ELECTRICITY RATES 154
APPENDIX F NATURAL GAS PRICING TARIFF FROM EGOLI GAS 155
APPENDIX G WITS UNIVERSITY USAGE AND BILLING TRENDS 156
APPENDIX H SPECIFIC COSTS 158
APPENDIX I PERSPECTIVE MODEL 159
APPENDIX J REFERENCE PLANT RESULTS 160
APPENDIX K MODEL INPUT 162
APPENDIX L COMPARISON MODEL 170
x
LIST OF FIGURES
Figure 2.1: The World's Solar Resource (Stine and Geyer, 2008) 8
Figure 2.2: Annual DNI Data for South Africa (NREL, 2008) 8
Figure 2.3: Parabolic Trough CSP Plant in the Mojave Desert (Sitenet, 2008) 10
Figure 2.4: Parabolic Trough and Power Plant of SEGS Type (Beerbauma and
Weinrebeb, 2000) 10
Figure 2.5: CLFR System (Power Technology, 2009) 11
Figure 2.6: Schematic Diagram Showing Interleaving Mirrors of the CLFR Collectors
(Mills and Morrison, 2000) 11
Figure 2.7: Central Receiver Plant (CSP, 2008) 13
Figure 2.8: Central Receiver, PHOEBUS Schematic (Beerbauma and Weinrebeb,
2000) 14
Figure 2.9: Dish-Stirling System (Pitz-Paal et al., 2005) 14
Figure 2.10: Dish Stirling System of a Schlaich Bergerman 10 kW (Beerbauma and
Weinrebeb, 2000) 15
Figure 2.11: Solar Chimney Technology (Beerbauma and Weinrebeb, 2005) 16
Figure 2.12: Delivery Period Extension (Geyer, 1999) 17
Figure 2.13: Power Booster and Fuel Saver in Hybrid Alternatives (Kolb, 1998) 19
Figure 2.14: Schematic of an ISCCS System (Hosseini et al., 2005) 20
Figure 2.15: S&L Cost Reduction Potential of CSP (S&L, 2003) 29
Figure 2.16: Methodology for the Ecostar Cost Study (Pitz-Paal et al., 2005) 31
Figure 2.17: Ecostar Cost Reduction Potential (Pitz-Paal et al., 2005) 32
Figure 3.1: Solar to Electric Efficiency 52
Figure 4.1: Financial Results 59
Figure 4.2: Numerical Evaluation Matrix 65
Figure 4.3: Cause and Effect Graph 65
Figure 7.1: Reference Plant Areas (120 kW(e)) 100
Figure 7.2: Plant Area for CLFR, ORC technologies (480 kW(e)) 100
Figure 7.3: Economic Results (120 kW(e) reference systems) 102
xi
Figure 7.4: Hourly Energy Flow for CLFR, ORC (480 kW(e)) 105
Figure 7.5: Average Energy Flow for CLFR, ORC (480 kW(e)) 105
Figure 7.6: Hourly Energy Flow for CLFR, ORC, with Storage (480 kW(e)) 105
Figure 7.7: Average Energy Flow for CLFR, ORC, with Storage (480 kW(e)) 105
Figure 7.8: Hourly Energy Flow for CLFR, ORC, with Hybridisation (480 kW(e)) 105
Figure 7.9: Average Energy Flow for CLFR, ORC, with Hybridisation (480 kW(e)) 105
Figure 7.10: West Campus Power Usage - CLFR, ORC (480 kW(e)) 106
Figure 7.11: West Campus Power Usage - CLFR, ORC (480 kW(e)) 106
Figure 7.12: West Campus Power Usage - CLFR, ORC, with Storage (480 kW(e)) 106
Figure 7.13: West Campus Power Usage - CLFR, ORC, with Storage (480 kW(e)) 106
Figure 7.14: West Campus Power Usage - CLFR, ORC, with Hybridisation (480
kW(e)) 106
Figure 7.15: West Campus Power Usage - CLFR, ORC, with Hybridisation (480
kW(e)) 106
Figure 7.16: LEC Results for CLFR, ORC technologies (480 kW(e)) 111
Figure 7.17: Payback Results for CLFR, ORC Technologies (480 kW(e)) 112
Figure A1: Schematic Flow Diagram of the SEGS 1 plant 140
Figure C1: Average Daily Data - JHB 151
Figure C2: Monthly Statistics-JHB 151
Figure C3: Hourly DNI Data - June- JHB 152
Figure C4: Hourly DNI Data - January- JHB 152
Figure G1: West Campus Usage- November 2007 156
Figure G2: West Campus Usage- June 2008 156
Figure G3: Monthly bill June 2008 157
Figure G4: Historical Billing Trend for West Campus 157
Figure J1: LEC for 120 kW(e) Reference Plants 161
Figure J2: Payback for 120 kW(e) Reference Plants 161
xii
LIST OF TABLES
Table 2-1: Advantages and Disadvantages of CSP 24

Table 3-1: Chemical Engineering Plant Cost Index 47
Table 3-2: CEPCI Average 48
Table 4-1: Perspective Model 67
Table 5-1: Potential Site Areas 74
Table 5-2: Optical Characteristics of the Parabolic Trough System 79
Table 5-3: Convection Losses for Parabolic Trough Collectors 80
Table 5-4: Radiation losses for Parabolic Trough Collectors 80
Table 5-5: Optical Characteristics of the Linear Fresnel System 81
Table 5-6: Convection Losses for CLFR Collectors 82
Table 5-7: Radiation Losses for CLFR Collectors 82
Table 5-8: Power Block Design Specifications 83
Table 5-9: Parabolic Trough Efficiencies 83
Table 5-10: Design Efficiencies 84
Table 5-11: Exchange Rates 85
Table 5-12: Economic Assumptions 88
Table 5-13: Average Economic Data 89
Table 5-14: International CSP REFITs (Geyer, 2007) 91
Table 7-1: Thermal Energy Flow 99
Table 7-2: Aperture Areas 99
Table 7-3: Economic Results 101
Table 7-4: Energy Results using CLFR, ORC 108
Table 7-5: Energy Results using CLFR, ORC with Storage 108
Table 7-6: Energy Results using CLFR, ORC with Hybridisation 109
Table 7-7: Billing Results using Normal CLFR, ORC 109
Table 7-8: Billing Results CLFR, ORC with Storage 110
Table 7-9: Billing Results using CLFR, ORC with Hybridisation 110
Table 7-10: Summary for CLFR, ORC Technologies at 480 kW(e) 111
xiii
Table 9-1: Summary (480 kW(e) systems) 128
Table B1 Summary of Evaluated Technologies 147
Table C1: Average Hourly Statistics for Direct Normal Solar Radiation Wh/m² 150
Table D1: Parabolic Trough and CLFR Verification 153
Table H1: Specific Costs 158
Table I1: Perspective Model for Twelve Alternatives 159
Table J1: Reference Plant Results 160
xiv
LIST OF SYMBOLS AND ABBREVIATIONS
Symbol Quantity Unit
Aa aperture area of solar field m2
ASf aperture area of the solar field m2
BOP balance of plant -

CDM Clean Development Mechanism -
CEPCI chemical engineering plant cost index -
CER Certified Emission Reduction -
CFsolar solar capacity factor -
CLFR compound linear Fresnel reflector -

CSP concentrating solar power -
C concentration factor -
DI direct normal irradiation kWh/m2/a
Enet net electricity generated kWh
Esolar net solar electricity generated kWh
ECOSTAR European concentrated solar thermal road mapping -

EPW energy plus weather -
fcr fixed charge rate -
FV future value R/$/€
GCR ground cover ratio -
GEF Global Environment Fund -
IB issuing body -
IEA International Energy Agency -
ISCCS integrated solar combined cycle system -
HVAC heating, ventilating and air conditioning -
kd real debt rate -
kinsur annual insurance rate -
K fuel annual fuel costs R/$/€
xv
K invest total capital investment R/$/€
KO&M annual operating and investment costs R/$/€
LEC Levelised Cost of Electricity (R/$/€)/kWh

MTPP modular trough power plant -
n life of plant/discount period years
NPO non-profit organisation -
NREL National Renewable Energy Laboratory (USA) -
ORC organic Rankine cycle -
Pnet net power output kW
PPA power purchase agreement -

PSA Plataforma Solar de Almeria -
PV present value R/$/€
Qɺ Sf thermal energy delivered by the solar field kW
Qɺ therm thermal energy input to the power cycle kW
r interest rate -
REFIT renewable energy feed-in tariff -
SANTRECT South African National Tradable Renewable Energy
Certificate Team -
SEGS solar energy generating systems
S&L Sargent and Lundy -
TA mean surface temperature of the absorber tube K
Tamb ambient temperature K
TES thermal energy storage -

TMY2 typical meteorological year -
TOU time of use -
U convection loss heat transfer coefficient W/m²K
Wits Witwatersrand (University) -
Wdesign net design output of the power block kW
xvi
Greek Symbols
Symbol Quantity Unit
α coefficient of absorption of the absorber tube -
γ mirror quality factor -
ε coefficient of emission of the absorber tube -
η par efficiency due to pumping parasitic losses -
η piping piping efficiency -
η pbnet net power block efficiency -
ηopt optical efficiency -
ηrec / pip receiver/piping efficiency -
η s −e net solar to electric efficiency -
η Sf solar field efficiency -
η stor storage efficiency -
ξ geo geometric Efficiency -
ξ IAM the incident angle modifier -
ξS shading losses within the solar field -
ξE intercept factor -
ξcos cosine losses -
ρ reflectivity of the mirrors -

σ Stefan-Boltzmann constant W/m-²K-4
τ1 transmission factor of the mirror glass cover -
τ2 transmission factor of absorber tube -
xvii
1 INTRODUCTION
1.1 Background
The prospects of climate change and, eventually, fossil fuel depletion, trigger a
growing interest in renewable energies in general. The benefits of renewable energy
systems were clearly defined in a political declaration agreed upon by government
representatives of 154 nations at the international “Renewables 2004” conference
held in Bonn, June 2004 as a follow-up to the 2001 World Summit on Sustainable
Development, Johannesburg. Benefits outlined included energy supply security,
equity and development, improved health, overcoming peak oil price fluctuations,
provision of clean water, close association with energy efficiency measures, climate
change mitigation, and the common belief that “there will be no need for war over
solar energy” (Philibert, 2005).
The use of renewable energy in the world has been implemented for many different
reasons. There is a huge drive for renewable energy in Europe mainly because of the
focus on reducing emissions and climate change mitigation. South Africa is well
endowed with renewable energy resources that can be sustainable alternatives to
fossil-fuels, so far these have remained largely untapped. South Africa released a
White Paper on Renewable Energy (DME, 2003) where it identified a heavy reliance
on coal to meet its energy needs mainly because it has a huge coal resource.
However, at the same time South Africa recognises that the emissions of greenhouse
gases, such as carbon dioxide, from the use of fossil fuels such as coal and petroleum
products has led to increasing concerns worldwide, about global climate change.
The driving force for energy security can be tackled through the diversification of
South Africa’s supply. The South African economy, which is highly dependent on
income generated from the production, processing, export and consumption of coal,
is vulnerable to the possible climate change response measures implemented or to
1
be implemented by developed countries. At the same time there are now increased
opportunities for energy trade. “Given increased opportunities for energy trade,
particularly within the Southern African region, Government will pursue energy
security by encouraging diversity of both supply sources and primary energy carriers. ”
(DME, 1998)
For this purpose, the Government will develop the framework within which the
renewable energy industry can operate, grow, and contribute positively to the South
African economy and to the global environment.
1.2 Motivation
From the background of renewable energy above, three major factors motivating
the use of renewable energy have arisen. These are:
• Economic reasons
• Energy security
• Climate change mitigation.
As a result of insufficient electrical power generation infrastructure investment in

South Africa in the last two decades compared to economic growth, power outages
have been experienced in South Africa since late 2007. This is having a detrimental
effect on South Africa’s economy and the need for energy security amongst
businesses and institutions has arisen.
Electricity production from fossil fuels, particularly coal, is a large contributor to the
CO2 burden. In South Africa some 90% of electricity production is by coal-fired
power stations and 30% of liquid fuels are derived from coal via the Fisher-Tropsch
process (Roos, 2009). In fact, the Sasolburg Secunda plant is the world’s largest
point source of CO2. Recognising this need for renewable energy, this study
2
investigates options to replace electricity production from coal with a renewable
source.
The University of the Witwatersrand, Johannesburg, (Wits) is experiencing heavy

electrical usage. In parallel with an energy usage and a consumption study to
understand usage patterns with a view of an energy efficiency strategy, a study was
envisioned to investigate alternative energy generation. It is important to note that
this study does not aim to provide a solution to financial distress but as in the case
of any study, the financial feasibility cannot be ignored and will play a very
important role in any implementation decisions.
For these reasons as well as the fact that Johannesburg has a high solar resource (as
opposed to other renewable resources - see Section 2.2.2), researchers at Wits
University have expressed interest in concentrating solar power (CSP). This study
investigates potential distributed power generation solutions for urban areas, with
Wits University’s West Campus as a case study.
1.3 Objectives
Several studies assessing the feasibility of CSP technologies have been performed
but they mainly emphasize large generating stations where land issues are
unimportant and can make use of the economies of scale to drive down the
Levelised Cost of Energy (LEC). The aim of this report is to review the use and the
implementation of several solar-thermal electric technologies in urban
environments and to carry out a technical and economic feasibility study applicable
to Johannesburg.
There are many benefits regarding the use of renewable energy. Currently costs are
certainly not one of these and it will also be part of this study to review these
benefits to the University of the Witwatersrand (Wits) by comparing them to
3
current sources of energy. Because of the expressed need and interest, this report
could possibly lead to the implementation of some form of solar-thermal technology
at Wits University.
Specific objectives identified include:
• To deliver a review of the relevant literature with regard to the development

of Solar Thermal power generation.
• To draw a comparison of several of the available technologies outlining

specifically what would be suitable in different applications. (e.g. off-grid, on-
grid, hybridisation, scaling effects etc)
• To report on the potential for solar thermal technologies in Johannesburg,

based on local conditions.
• To perform a technology screening in order to select the system that will best
suit implementation at the University of the Witwatersrand.
• To identify suitable technologies and develop a model/conceptual design

configurations of possible CSP generating systems for Wits University. This
will explore themes which will include:
o Technical viability - This will show which of the technologies are

suitable in terms of functional criteria such as space usage,
modularity, maturity of technology etc. It will also assess the
suitability of Johannesburg’s solar resource with respect to CSP
generation.
o Economic/financial viability - This model will explore different
options, showing the financial feasibility in the implementation of the
technology, from the equipment costs to the actual running and
electricity costs, as well as the savings experienced in different cases.
4
2 LITERATURE REVIEW
2.1 Utility and Distributed Generation
Utility Scale plants are usually large centralised facilities, such as traditional coal
fired plants, which can reach generation capacities of thousands of MWs. These
plants have excellent economies of scale, but usually transmit electricity long
distances. Most of these plants are built this way due to a number of economic,
health and safety, logistical, environmental, geographical and geological factors. For
example, coal power plants are built away from cities to prevent their heavy air
pollution from affecting the populace; in addition such plants are often built near
collieries to minimize the cost of transporting coal.
Distributed generation reduces the amount of energy lost in transmitting electricity

because the electricity is generated near where it is used, perhaps even in the same
building. Distributed energy resource systems are small-scale power generation
technologies (typically in the range of 3 kW to 10000 kW) used to provide an
alternative to or an enhancement of the traditional electric power system (IEEE,
2005).
A report prepared by Hoff (2000) discusses how local governments benefit from
distributed resources. Such benefits include:
• Improving the environment

• Guiding economic development
• Ensuring electrical system reliability for constituents
• Providing constituents with energy security
• Providing disaster relief support.
5
2.2 CSP Technology: Basic concepts
2.2.1 Introduction
Concentrating solar power technologies (CSP) only use solar beam radiation as
opposed to diffuse solar radiation, concentrating it several times to reach higher
energy densities - and thus higher temperatures when the radiation is absorbed by
some material surface. The conversion of this heat into mechanical energy is done
using similar processes to conventional power cycles, for example the Rankine cycle,
converting heat from burning coal into electricity.
There is a variety of technologies that are available, for example, the Californian 354
MW parabolic trough solar electric generating systems which have been operating
for more than 20 years (Pitz-Paal et al., 2005). The major deterrent for solar
electricity generation is the relatively high specific investment cost of the solar
collector systems.
Concentrating solar power plants offer a very promising option for a sustainable
electricity supply. Solar energy, as a source, fluctuates naturally, first as a result of
diurnal cycles and secondly as a result of cloud passage, leading to fluctuations in
generation. This has led to various technologies that have been developed to solve
this intermittency. Because it uses a thermal phase, CSP technologies can easily
make power production firm and even dispatchable, either by storing the heat in
various forms, or by backing its production by some fossil fuel burning – in both
cases using the same steam turbines and generators. Other technologies such as
wind power that do not convert thermal energy into electricity can also implement
storage but at a higher cost because the price of storing electricity is much higher
than storing thermal energy (Pitz-Paal et al., 2005).
CSP technologies are best suited to areas with high direct solar radiation. According
to Solel (ISRAEL21c, 2007), a solar thermal plant built on just one percent of the
6
surface of the Sahara Desert could provide the entire world's electricity demands.
These areas are widespread, but not universally found over the globe.
Recognising both the environmental and climatic hazards to be faced in the coming
decades and the continued depletion of the world‘s most valuable fossil energy
resources, concentrating solar thermal power can provide critical solutions to global
energy problems within a relatively short time frame and is capable of contributing
substantially to carbon dioxide reduction efforts. Among all the renewable
technologies available for large-scale power production today and for the next few
decades, CSP is one with the potential to make major contributions of clean energy
because of its relatively conventional technology and ease of scale-up.
2.2.2 Solar Energy Resource

Before introducing the CSP technologies, the solar resource requirement is defined.
Solar thermal power can only use direct sunlight, called ‘beam radiation’ or Direct
Normal Irradiation (DNI), i.e. that fraction of sunlight which is not deviated by
clouds, fumes or dust in the atmosphere and that reaches the earth’s surface in
parallel beams for concentration. Hence, it must be sited in regions with high direct
solar radiation. Suitable sites should receive at least 1700 kilowatt hours (kWh) of
sunlight radiation per m2 annually (Stine and Geyer, 2008), whilst best site locations
receive more than 2800 kWh/m2/year. Typical site regions, where the climate and
vegetation do not produce high levels of atmospheric humidity, dust and fumes,
include steppes, bush, savannas, semi-deserts and true deserts, ideally located
within less than 40 degrees of latitude north or south. Therefore, the most
promising areas of the world include the South-Western United States, Central and
South America, North and Southern Africa, the Mediterranean countries of Europe,
the Near and Middle East, Iran and the desert plains of India, Pakistan, the former
Soviet Union, China and Australia (Stine and Geyer, 2008). This is shown in Figure
2.1.
7
Figure 2.1: The World's Solar Resource (Stine and Geyer, 2008)
When it comes to siting in urban areas, it is important to bear in mind that there
may be different design considerations such as the fact that structures found in
urban areas such as buildings and towers may cast shadows onto the catchment
area which may be on a field or even on top of other buildings. It will be important
to consider each situation. It will be assumed that the solar data collected will be
completely available at the chosen site.
Figure 2.2: Annual DNI Data for South Africa (NREL, 2008)
8
Figure 2.2 shows the DNI data with 40 km2 sensitivity. Accordingly the DNI for the
Johannesburg region is typically between 5.0 - 6.0 kWh/m2/day which equates to
1825-2190 kWh/m2/year.
2.3 Collector Types
The solar thermal technologies to be evaluated in this study vary, but most can be
classified into the following broader categories:
• Line Focussing Systems
o Trough Technology
o Linear Fresnel Collectors.
• Point Focussing Systems
o Central Receiver Technology
o Dish-Stirling.
• Non-Concentrating type
o Solar Chimney.
2.3.1 Parabolic Trough Collector System

Parabolic trough power plants are line-focusing CSP plants. Trough systems use the
mirrored surface of a linear parabolic concentrator to focus direct solar radiation on
an absorber pipe running along the focal line of the parabola (Figure 2.3). The heat
transfer fluid inside the absorber pipe is heated and pumped to the steam generator,
which, in turn, is connected to a steam turbine (STI, 2005) (Shown in Figure 2.4).
9
Figure 2.3: Parabolic Trough CSP Plant in the Mojave Desert (Sitenet, 2008)
Figure 2.4: Parabolic Trough and Power Plant of SEGS Type (Beerbauma and Weinrebeb, 2000)
10
2.3.2 Compound Linear Fresnel Reflector (CLFR)
In the CLFR configuration, large fields of modular Fresnel reflectors concentrate
beam radiation to a stationary receiver several metres high. This receiver contains a
second stage reflector that directs all incoming rays to a tubular absorber (Häberle
et al., 2002).
Figure 2.5: CLFR System (Power Technology, 2009)
Mills and Morrison (2000) describe an advanced CLFR technology, noting several
technological aspects that need to be developed further. This concept includes a
secondary reflector, installed to help direct the insolation onto the absorber. The
advantage of this system is that it allows for densely packed arrays, because
patterns of alternating reflector inclination can be set up such that the closely
packed reflectors can be positioned without shading and blocking. The ‘interleaving’
of mirrors between two linear absorber lines is shown in Figure 2.6.
Figure 2.6: Schematic Diagram Showing Interleaving Mirrors of the CLFR Collectors (Mills and
Morrison, 2000)
11
This arrangement minimizes beam blocking between adjacent reflectors and allows
higher reflector densities and lower absorber tower heights to be used. Available
area can be restricted in industrial or urban situations. Avoidance of large reflector
spacing and high towers are an important cost issue when one considers the cost of
ground preparation, array structure and tower structure. Using the CFLR reflectors
for generation of steam, however, offers no obvious form of thermal storage
(Section 2.4.1), only offering generation during sunlight hours.
The CLFR power plant, designed by Mills and Morrison (2000), includes the
following additional features which enhance the system cost/performance ratio.
Points a) and b) being unique to this design.
a) The array uses flat or elastically curved reflectors instead of costly sagged
glass reflectors. The reflectors are mounted close to the ground, minimising
structural requirements.
b) The heat transfer loop is separated from the reflector field and is fixed in
space thus avoiding the high cost of flexible high pressure lines or high
pressure rotating joints as required in the trough and dish concepts.
c) The heat transfer fluid is water, and passive direct boiling heat transfer
could be used to avoid parasitic pumping losses and the use of expensive
flow controllers. Steam supply may either be direct to the power plant steam
drum, or via a heat exchanger.
d) All-glass evacuated tubes with very low radiative losses can be used as the
core element of the linear absorber array.
e) Maintenance will be lower than in other types of solar concentrators
because of nearly flat reflectors and ease of access for cleaning, and because
the single ended evacuated tubes can be removed without breaking the heat
transfer fluid circuit.
12
2.3.3 Central Receiver Technologies
A circular array of heliostats (large individually tracking mirrors) is used to
concentrate sunlight on to a central receiver mounted at the top of a tower. A heat-
transfer medium in this central receiver absorbs the highly concentrated radiation
reflected by the heliostats and this thermal energy is be used for the subsequent
generation of electricity in a Rankine or Brayton cycle turbine (Figure 2.8). To date,
the heat transfer media demonstrated includes water/steam, molten salts, liquid
sodium and air. If pressurised gas or air is used at very high temperatures of about
1,000°C or more as the heat transfer medium, it can even be used to directly replace
natural gas burning in a gas turbine, thus making use of the excellent cycle efficiency
(60% and more) of modern gas and steam combined cycles (STI, 2005). Such a
system is shown below in Figure 2.7.
Figure 2.7: Central Receiver Plant (CSP, 2008)
13
Figure 2.8: Central Receiver, PHOEBUS Schematic (Beerbauma and Weinrebeb, 2000)
2.3.4 Dish-Stirling Systems

A parabolic dish-shaped reflector is used to concentrate sunlight on to a receiver
located at the focal point of the dish. The concentrated beam radiation is absorbed
into the receiver to heat a fluid or gas (air) to approximately 750°C. This fluid or gas
is then used to generate electricity in a small piston or Stirling engine or a micro-
turbine, attached to the receiver. A photo and schematic of the Dish-Stirling system
is shown below in Figure 2.9 and Figure 2.10 respectively.
Figure 2.9: Dish-Stirling System (Pitz-Paal et al., 2005)
14
Figure 2.10: Dish Stirling System of a Schlaich Bergerman 10 kW (Beerbauma and Weinrebeb, 2000)
2.3.5 Solar Chimney Technology

The solar chimney consists of three essential elements: the solar collector, vertical
chimney and wind turbine. The solar collector consists of a transparent circular roof
which is open along the outside edge and situated near the ground. As the sun heats
the ground, it heats the air within the roof. The rise in air temperature as well as the
density decrease induces the heated air to rise through the vertical chimney in the
centre. This rising air turns a wind turbine to create electrical energy through the
conversion of kinetic energy. The warm rising air is constantly replaced by cool air
flowing in through the sides. Based on the test results, it was estimated that a 100
MW plant would require a 1000 m tower and a greenhouse of 20 km2 (Haaf et al.,
1983). A schematic of this system is shown below in Figure 2.11.
15
Figure 2.11: Solar Chimney Technology (Beerbauma and Weinrebeb, 2005)
2.4 Variations in Design and Common Technologies
2.4.1 Storage
Most renewable resources, including solar radiation, are intermittent in nature. A
distinct advantage of CSP plants compared with other renewable energies, such as
photovoltaic cells (PV) and wind, is the possibility of using relatively cheap storage
systems. That is, storing the thermal energy itself, a method which is financially
more feasible than storing electricity.
The principal options for using Thermal Energy Storage (TES) in a solar thermal
system highly depend on the daily and yearly variation of radiation and on the
electricity demand profile.
The main options, as identified by Pilkington Solar (2000), are:
• Buffering
• Delivery period displacement
• Delivery period extension.
The goal of a buffer is to smooth out transients in the solar input caused by passing
clouds, which can significantly affect operation of solar electric generating systems.
16
The efficiency of electrical production will degrade with intermittent insolation.
Buffer TES systems would typically require small storage capacities (maximum 1
hour full load).
Delivery period displacement requires the use of a larger storage capacity. The
storage shifts some or all of the energy collected during periods with sunshine to a
later period (possibly periods that have higher tariffs etc). This type of TES does not
necessarily increase either the solar fraction or the required collection area. The
typical size ranges from 3 to 6 hours of full load operation.
The size of a TES for delivery period extension will be of similar size (3 to 12 hours of
full load). However, the purpose is to extend the period of power plant operation
with solar energy. This TES increases the solar fraction and requires larger solar
fields than a system without storage. The operating model of such a system is given
in Figure 2.12 where additional thermal energy is collected during the day and is
utilised for electric generation after sun-set.
Figure 2.12: Delivery Period Extension (Geyer, 1999)
Design Criteria
A key issue in the design of a thermal energy storage system is its thermal capacity -
the amount of energy that it can store and provide. However selection of the
17
appropriate system depends on many cost-benefit considerations (Pilkington,
2000).
The cost of a TES system mainly depends on the following items:

• The storage material itself
• The heat exchanger for charging and discharging the system
• The cost for the space and/or enclosure for the TES
Pilkington outlines the most important design criteria as well as the crucial
technical requirements when choosing suitable storage technologies. Different
storage concepts are further discussed in Appendix A.
2.4.2 Hybrids
Hybrid systems, which make use of fossil fuels, are often used to make CSP
investments bankable. Solar energy can also be used to reduce fossil fuel usage
and/or boost the power output to the steam turbine (Kolb, 1998).
Typical daily power output from the hypothetical “power boost” hybrid power plant
is depicted in Figure 2.13. From the figure it can be seen that in a power boost
hybrid plant, a solar-only plant is “piggybacked” on top of a base-loaded fossil-
fuelled plant. In the power boost hybrid plant, additional electricity is produced by
over-sizing the steam turbine, contained within a coal-fired Rankine plant or the
bottoming portion of a combined-cycle plant, so that it can operate on both full fossil
and solar energy when solar is available. Studies of this concept have typically
oversized the steam turbine from 25% to 50% beyond what the turbine can produce
in the fossil-only mode (Kolb, 1998). Over-sizing beyond this range is not
recommended because the thermal-to-electric conversion efficiency will degrade at
the part loads associated with operating in the fuel-only mode. This over sizing of
the steam turbine has been typically proposed for many of the World Bank and
Global Environment studies where they would make use of parabolic troughs as the
solar collector in the ISCCS proposals (see Section 2.4.3).
18
In the “fuel saver” plant, the fuel usage is reduced when solar energy is available and
electricity output is constant. In a Rankine-cycle application, the solar steam
generator can be sized to provide the entire input to the steam turbine or a
fractional amount. When hybridising, it is preferred to contribute a fractional
amount of heat from solar. This keeps the fossil boiler hot all the time and prevents
daily start-up losses and thermal cycles. The Solgate study that uses high
temperature volumetric air in the receiver of the Central receiver makes use of the
“fuel saver” principle (Pitz-Paal et al., 2005).
Figure 2.13: Power Booster and Fuel Saver in Hybrid Alternatives (Kolb, 1998)
2.4.3 Integrated Solar Combined Cycle System (ISCCS)

The ISCCS configuration has been considered for a number of Global Environment
Fund (GEF) trough projects (World Bank, 2006). The ISCCS integrates solar steam
into the Rankine steam bottoming cycle of a combined-cycle power plant (Schematic
shown in Figure 2.14). The general concept is to oversize the steam turbine to
handle the increased steam capacity. At the high end, steam turbine capacity can be
approximately doubled, with solar heat being used for pre-heating and superheating
steam. Unfortunately when solar energy is not available, the steam turbine must run
at part load and thus reduced efficiency. Doubling the steam turbine capacity would
result in a 25% design point solar contribution. Because solar energy is only
available about 25% of the time, the annual solar contribution for trough plant
without thermal storage would only be about 10% for a base-load combined –cycle
plant (Price and Kearney, 2003).
19
Figure 2.14: Schematic of an ISCCS System (Hosseini et al., 2005)
2.4.4 Direct Steam Generation (DSG)

In the Direct Steam Generation (DSG) concept, steam is generated directly in the
parabolic-trough collectors. There is a reduction of costs found in the elimination of
traditional heat transfer fluid through the use of DSG (Price and Kearney, 1999).
This technology also reduces efficiency losses in the heat transfer process. DSG
should also improve the solar field operating efficiency due to lower average
operating temperatures and improved heat transfer in the collector receiver. The
trough collectors would require some modification due to the higher operating
pressure and lower fluid flow rates. Control of a DSG solar field is more complicated
than traditional systems and may require a more complex design layout and a tilted
collector. DSG also makes it more difficult to provide thermal storage. A pilot plant
was demonstrated at the Plataforma Solar de Almería (PSA) in Spain (Price and
Kearney, 1999).
20
2.4.5 Organic Rankine Cycles (ORC)
Traditionally, Organic Rankine Cycle Plants (ORC) are used for lower temperature
heat sources such as geothermal or waste heat recovery. The low resource
temperature results in low efficiency of the ORCs; however, ORCs can be designed to
operate at substantially higher efficiencies with trough systems. ORCs use organic
(hydrocarbon) fluids that can be selected to best match the heat source and heat
sink temperatures (Prabhu, 2006).
ORCs operate at lower temperatures than steam Rankine systems and thus can
reduce trough operating temperatures from 390 ˚C to 304 ˚C. This means that an
inexpensive heat transfer fluid such as Caloria may be used instead of the existing
fluid. Since Caloria is inexpensive, it can be used in a simple two-tank thermal
storage system similar to the thermal storage system at the SEGS I plants in the
Mojave Desert. Lower solar field operating temperatures are likely to translate into
lower capital cost and more efficient solar field equipment (Prabhu, 2006).
If a water resource is scarce ORCs can also be designed to use air-cooling for the
power cycle (as can be done for other cycles). This and the fact that the power cycle
uses a hydrocarbon for a working fluid (instead of steam) means that the plant
needs virtually no water to operate. Water consumption is reduced by 98%. Mirror
washing is only about 1.5% of the water use at the SEGS plants, meaning that the
water contribution to cleaning will be minimal (Prabhu, 2006).
These plants are capable of automatic start-up, safe shutdown, and regulation with
varying solar conditions. Because of their simplicity they can generally be operated
remotely. This helps to reduce operating and maintenance (O&M) costs which have
been one of the key reasons for concentrated solar power (CSP) technologies to
increase in size.
21
ORC systems have a number of disadvantages as well. ORC systems generally have
lower efficiencies than steam cycles that run at higher temperatures and pressures.
However, the efficient steam cycles come at the price of more capital investment and
the need for higher resource temperatures. The use of air-cooling means that ORC
cycles are negatively impacted by high ambient temperatures (Prabhu, 2006).
2.5 Data Sources
As already stated, several studies on the feasibility of the use of CSP generation have
been performed, ranging from technology-specific to purely economic comparisons.
The majority of these detailed reports have come from large organisations such as
the National Renewable Energy Laboratory (NREL) in the USA. The following is a list
of some of these more detailed reports, each listed with the year of respective
publication.
• Solar Electric Generation – A Comparative Overview, 1997

• Eskom CSP Pre-feasability Study, 2001
• Modular Trough Power Plants (MTPP), 2001
• Solarmundo line focussing Fresnel collector, 2002
• Assessment of Parabolic Trough and Power Tower Solar Technology, 2003
• European Concentrated Solar Thermal Road-Mapping Report, 2005
• The Present And Future Use Of Solar Thermal Energy, 2005
• California studies for NREL, 2005, 2006
• Assessment of the World Bank Group/GEF Strategy, 2006.
22
2.5.1 Solar Electric Generation – A Comparative Overview (1997)
Trieb et al. (1997) conducted a comparative review of different technologies, costs
and environmental impacts of solar electricity generation. The study shows that the
different approaches cover a wide range from units producing a few Watts to utility-
scale plants and from isolated to grid-connected systems.
Trieb et al. also identified two technical solutions to address the many drawbacks of
solar thermal technology. The first solution is the hybridisation of solar power plants
with fossil back-up systems. A fossil back-up system will allow for the compensation
of solar input fluctuations and permits night-time operation increasing the total
capacity factor. The second solution is the integration of energy storage systems into
the solar plant. This will also allow for the compensation of solar input fluctuations
with storage being possible for as long as 12 hours. This, however, does increase the
solar multiple and increase the capital costs of the system quite significantly. (The
solar multiple is the size of solar field relative to a field providing 100% design
power at peak collection times. This means a solar multiple of 1.2 represents a field
that delivers 20% more energy at solar noon than is required by the heat engine
generator).
Trieb et al. identified several advantages and disadvantages of the various

technologies and these have been listed below.
23
Table 2-1: Advantages and Disadvantages of CSP
Advantages Disadvantages
• Stand-alone units
• Very high concentration ratios, working
• Low power availability and few annual full load
temperatures and efficiencies
hours
Dish-Stirling
• Long term experience with small scale power

• Requires rigid support structures and perfect
plants and single units.
tracking that leads to high costs
• Options for distributed as well as centralised
• No experience with large-scale utility scale systems
electricity supply systems
• Water requirement for cleaning.
• Modularity of the system, benefits of mass
production, no scale restriction
• Simple operation and maintenance.
• The glass collector uses diffuse and beam • Very low solar to electric conversion efficiency
radiation. • Hybridisation not possible
• The soil under the collector acts as heat storage, • Equivalent full load hours restricted to
Solar Chimney
avoiding sharp fluctuations and allowing power approximately 2500h/a.

supply after sunset. • Large completely flat areas required for the
• Easily available and low cost materials for collector
construction • The high tower needed results in a large material
• Simple, fully automatic operation requirement for the system.
• No water requirements.
• High solar efficiencies • The solar energy and the fossil backup fuel are
• High steam temperatures converted to electricity with relatively low steam
Central Receiver
• Simple hybridisation with fuel oil or natural gas. cycle efficiency.

• Modular solar components (heliostats) with high • Heliostats require very stable supports for the
mass production potential mirrors and two axis tracking.
• Simple operation strategy • Water needed for mirror cleaning.
• Process steam generation for eventual • They are suited mainly for large scale electricity
cogeneration. generation.
24
2.5.2 Eskom CSP Pre-feasibility Study (2001)
In 2001 Eskom performed a state-of-the-art review of CSP technologies with the
goal of implementing a utility scale power plant in South Africa. Several technology
options were considered, while detailed design evaluations of a Central Receiver
and Parabolic Trough system were considered. A full economic study as well as an
environmental assessment for the Northern Cape was performed (van Heerden,
2001). This study was co-funded by the Global Environment Fund as well as the
World Bank. It is not publicly available information but has been provided by the
CSIR who has been given rights to it from the World Bank.
The study comprised three tasks; these have been identified as follows:
• The identification of fourteen different CSP technologies and their design
variations. Information was compiled from the published literature and
demonstration and operational plants where available.
• The second task involved the compilation of Typical Meteorological Year data
(TMY) for the reference site in Upington as well as a full strategic
environmental assessment for the Northern Cape Province.
• The third task involved the development of a simulation model that would
predict the performance of two selected technologies each at 100 MW(e). A
full economic assessment and optimisation was performed on these
technologies.
Eskom concluded that the central receiver technologies and parabolic trough
technologies, at the time of writing, have equivalent competitiveness. The central
receiver technologies, however, offered the greatest potential for cost reductions in
the future. By introducing a 100 MW(e) pilot plant in Upington, Eskom would be
able to produce the cheapest solar electricity in the world. CSP generated power will
be more costly than coal power for the foreseeable future but still remains an
attractive electricity source, primarily for peaking power production, because of its
environmental benefits.
25
2.5.3 Modular Trough Power Plants (MTPP) (2001)
In their paper, Hassani and Price, (2001) recognized that a number of factors are
creating an increased market potential for small trough power technology. They
conducted research into the feasibility of Modular Trough Power Plants (MTPP).
The reasons for conducting the research are as follows:
• There is a need for distributed power systems for rural communities

worldwide.
• The need to generate more electricity by non-combustion renewable
processes.
• The need for sustainable power for economic growth in developing
countries.
• The deregulation and privatisation of the electrical generation sector
worldwide.
Hassani and Price concluded that the ORC power cycles and parabolic trough solar
collector technology have been successfully demonstrated separately. With the
current state of these technologies, the modular trough power plant is a
technologically viable concept. Their analysis indicates that cycle efficiencies in the
range of 23% for a solar resource temperature of 580 ˚F (304 ˚C) are possible. Their
analysis was performed using meteorological data in Barstow, California with a net
electric capacity of the plant being 1 MW(e). Using cost and performance
assumptions outlined in their report, a cost of power around $0.20/kWh (2001)
appears to be feasible.
2.5.4 Solarmundo line focussing Fresnel collector (2003)

The Belgian company Solarmundo claim that its Fresnel collector is more cost
effective than existing CSP-systems. Solarmundo operates a 2500 m² prototype in
Liège, Belgium (Häberle et al., 2002).
26
In their paper Häberle et al. present optical and thermal properties of the
Solarmundo collector, which were calculated using ray-tracing and computational
fluid dynamics simulations. It is the basis for a simulation model to calculate the
thermal output of the collector for different sites. The behaviour of Fresnel
collectors compared to parabolic troughs is also discussed. An outlook on the
achievable costs of electricity is given.
2.5.5 Assessment of Parabolic Trough and Power Tower Technology

(2003)
NREL has produced an assessment of parabolic trough and central receiver costs
and performance forecasts. The assessment was performed by the consulting group
Sargent and Lundy LLC (S&L, 2003).
The following are specific themes that Sargent and Lundy investigated:
• The examination of the current trough and tower baseline technologies that
are examples of the next plants to be built, including a detailed assessment of
the cost and performance basis for these plants.
• Analysis of the industry projections for technology improvement and plant
scale-up to 2020, including a detailed assessment of the cost and
performance projections for future trough and tower plants based on factors
such as technology R&D progress, economies of scale, economies of learning
resulting from increased deployment, and experience-related O&M cost
reductions resulting from deployments.
• Assessment of the level of cost reductions and performance improvements
that, based on Sargent and Lundy experience, are most likely to be achieved,
and a financial analysis of the cost of electricity from such future solar trough
and tower plants.
27
Sargent and Lundy concluded that CSP is a proven technology for energy
production, and that significant cost reductions are achievable assuming that
reasonable deployment of CSP technologies occurs. Sargent and Lundy
independently projected capital and operating and maintenance costs, from which
the levelised energy costs were derived, based on a conservative approach whereby
the technology improvements are limited to current demonstrated or tested
improvements.
In their report Sargent and Lundy identified several market barriers that need to be
overcome to aid the implementation of CSP technologies in bulk scale generation
facilities. For CSP technologies to reach market acceptance the following market
entry barriers need to be overcome:
• Market expansion of trough and tower technology will require incentives to

reach market acceptance (competitiveness). Both tower and trough
technology currently produce electricity that is more expensive than
conventional fossil-fuelled technology. Analysis of incentives required to
reach market acceptance was not within the scope of the report.
• Significant cost reductions will be required to reach market acceptance
(competitiveness). Sargent and Lundy focused on the potential of cost
reductions with the assumption that incentives will occur to support
deployment through market expansion.
They also concluded that cost reductions are achievable for CSP systems, assuming
reasonable deployment occurs. They predicted projected energy costs, reductions
and performance improvements for the long term (2020). These are summarised in
Figure 2.15. This figure describes cost reductions with time. This is based in turn on
cost reductions with numbers of units in the field. The Sunlab study referred to in
Figure 2.15 forms the basis for comparison.
28
Figure 2.15: S&L Cost Reduction Potential of CSP (S&L, 2003)
To arrive at a credible prediction, Sargent and Lundy went into great detail outlining
all the technologies involved in each system, comparing them as well as costs to
various existing systems and predictions made by other organisations, for example
the SunLab cost model, outlining the differences in methodologies as well as results.
No marketing analysis was performed in terms of the power generation market and
its associated issues. Included in such an analysis would be the required incentives
needed for effective deployment.
The Sargent and Lundy report did not include a bottom-up cost estimate. Instead,
Sargent and Lundy drew heavily from industry experience, vendor quotes, and other
sources rather than recreate all this analysis on its own. The methodology used by
Sargent and Lundy stands on its own as a credible assessment of the status and
potential of parabolic trough and central receiver technologies. The results obtained
in the Sargent and Lundy study are insufficient for the current study because of the
scaling differences and other assumptions used.
29
The appendices in the Sargent and Lundy report are quite extensive and detail
methods of calculation for different aspects of the feasibility. These have been used
in this report, for example the same scaling methods were followed. Also the
equations used in the calculation of the solar capacity factor, field area, LEC and
availability were used in this report.
2.5.6 European Concentrated Solar Thermal Road-Mapping Report

(2005)
The European Concentrated Solar Thermal Road-Mapping Report (ECOSTAR) is a
document prepared for the EU which compares major CSP technologies under the
following objectives (Pitz-Paal et al., 2005):
• To identify the potential European technical innovations with the highest

impact on CSP cost reduction.
• To focus the European research activities and the national research
programs of the partners involved on common goals and priorities.
• To broaden the basis of industrial and research excellence and to solve
multidisciplinary, CSP specific, problems.
The approach of the document was to analyse the impact on cost of different
innovations applied to a reference system in order to identify those with the highest
impacts. Cost and performance information of the reference systems used were at
different levels of maturity. The evaluation therefore focused on the identification of
the major cost reduction drivers for each of the considered reference systems and
identified the impact of technical innovation approaches. This led to
recommendation on R&D priorities as well as to recommendation on changes in the
political framework needed to achieve a successful deployment. The methodology
for the cost study is depicted in Figure 2.16.
30
Figure 2.16: Methodology for the Ecostar Cost Study (Pitz-Paal et al., 2005)
In their conclusion they identified different classes of innovation whereby

uncertainty is addressed by providing optimistic and pessimistic bounds on the
input data for the performance and cost model, resulting in appropriate bounds for
the LEC values and cost reduction percentages.
Many of the systems considered are planned for commercial deployment in Spain,
which at the time of reporting recently enacted an incentive of around 21
cents€/kWh for solar thermal electricity (Technologies found in Appendix B). The
present ECOSTAR evaluation estimates levelised electricity cost of 17-18
cents€/kWh for initial systems currently being built and some completed systems in
Spain. These cost estimates will probably deviate from electricity revenues needed
for the first commercial plants in Spain because they were evaluated using a
simplified methodology including the financing assumptions recommended by the
IEA (1991) for comparative studies like this.
The other technologies analyzed are currently planned in significantly smaller pilot
scales of up to 15 MW(e). The LEC is significantly higher for these small systems
ranging from 19 to 28 cents€/kWh. Assuming that several of the smaller systems
are built at the same site to achieve a power level of 50 MW and take benefit of a
similar O&M effort as the larger plants, LEC estimates of all of the systems also
31
range between 15 and 20 cents€/kWh. The systems achieve a solar capacity factor
of up to 30% under these conditions (depending on the availability of storage).
Figure 2.17 shows the cost reduction potential as predicted by the ECOSTAR
roadmap for the 7 CSP technologies investigated in the study based on the LEC for
the 50 MW(e) reference systems and assuming a combination of selected
innovations for each system.
Figure 2.17: Ecostar Cost Reduction Potential (Pitz-Paal et al., 2005)
Method used in Ecostar

Ecostar performed a major comparison between several of the existing CSP systems.
The goal of the Ecostar study is the comparison of different technical innovations,
therefore any project specific data (e.g. tax influences, or financing conditions) are
neglected. The approach is kept simple, but it is appropriate to perform the relative
comparison necessary to quantify the impact of different innovations.
The model uses common assumptions for the site, meteorological data and load
curve. The common assumptions used in the ECOSTAR model are as follows:
32
The site under analysis is Seville, Spain 5.9 ° W, 37.2° N, 20 m above sea level, land
costs 2,000,000 €/km². Meteorological data and Direct Normal Irradiance for Seville
are used. (DNI 2014 kWh/m²a; average Temp 19,5C°, Min = 4,1°C, Max = 41,4°C). It
is analysed in free-load operation or in hybrid operation with 100% load between
9:00 a.m. and 11:00 p.m. every day. An average availability of 96% to account for
forced and scheduled outages results in a capacity factor of 55%.
The Ecostar model calculates the annual electricity production hour by hour, taking
into account the instant solar radiation, load curve, part load performance of all
components (depending on load fraction and ambient temperature), operation of
thermal energy storage, and parasitic energy requirements.
The reference size of all systems is assumed to be 50 MW(e) net.
2.5.7 The Present and Future Use of Solar Thermal Energy (2005)
Philibert (2005) produced a report for the International Energy Agency (IEA) on the
present and future use of solar thermal energy. His review not only included the use
of CSP technologies but other solar thermal technologies such as the use of passive
solar architecture and the production of fuels which provides an interesting
discussion on the extent and possibilities of solar thermal applications.
2.5.8 California Studies for NREL (2005, 2006)

NREL also commissioned a project with the goal to evaluate the feasibility of
developing up to 1000 MW(e) of parabolic trough solar thermal power plants to
serve municipal utility electricity demand in the State of California. This was
presented to NREL by Solargenix Energy (Solargenix Energy, 2005).
33
The objectives as laid out in the document were to:
• Estimate the relative performance and cost in different regions of the state
• Examine siting issues for solar parabolic trough power plants.
• Identify specific permitting requirements, with an emphasis on unique issues
associated with this technology.
• Discuss technology options that include a reduction of cooling water
consumption and add a thermal storage capability to the plants.
• Explore financial and business models, and associated incentives, that might
lead to accelerated development and deployment in California.
• Formulate a draft power purchase agreement for use between an IPP
developer and a municipal utility.
The direct normal solar radiation in specific areas in southern California is large
enough to generate thousands of GW using CSP technology. Although currently
limited by transmission availability, this still represents a very large and attractive
resource for the California Municipalities. Trough technology is proven and
commercial, but its current cost makes selection difficult for the cost-conscious
Municipalities. When the added costs of future fuel price volatility and
environmental regulations are considered, the near-term costs of CSP appear close
to fossil-fuelled alternatives. Furthermore, the long-term trend suggests a crossover
between CSP and fossil-fuelled generation costs within about 5 to 10 years.
Another document that was prepared for NREL details the economic, energy, and
environmental benefits of concentrating solar power in California (Stoddard et al.,
2006). Emphasis was placed on in-state economic impact in terms of direct and
indirect employment created by the manufacture, installation, and operation of CSP
plants. The environmental impact of CSP relative to natural gas fuelled counterparts,
as well as the value of CSP as a hedge against natural gas price increases and
volatility, was studied.
34
2.5.9 Assessment of the World Bank Group/GEF Strategy (2006)
The World Bank (2006) produced a strategy report detailing the market
development of CSP technologies. This study presents an independent review of the
implementation progress for several World Bank/Global Environment Fund (GEF)
funded projects in the context of the long-term strategy for solar thermal
development. The study team undertook extensive consultations with stakeholders
and made some specific recommendations with regard to project implementation.
In particular, they emphasised the need for flexibility in technology choice and
implementation approach.
The World Bank and GEF undertook the development of four ISCCS solar thermal
projects in Mexico, Morocco, Egypt, and India. All four have experienced
implementation problems. These implementation problems arose mainly from three
specific issues, these being:
1) The contradiction between the drivers of economic development in

developing countries, i.e. poverty alleviation, and those of the developed
world, i.e. environmental concerns, generates a mismatch of global
expectations and local willingness to support these projects.
2) There has been insufficient dialogue between GEF and the CSP industry
during project design, adoption of the CSP strategy, and project
implementation.
3) GEF has remained the only significant funding source for these CSP plants.
The main objective of the assignment was to assess the strategy being followed by
the World Bank/GEF for solar thermal power technology in light of:
1. The current state of technology, costs, and market development.

2. The difficulties experienced by the GEF co-financed projects, assessing the
three primary risks facing the Bank/GEF portfolio.
35
a. Limited industry response.
b. Uncertainty of meeting the cost and performance targets.
c. Uncertainty of sustainability and replicability arising from the absence of
long-term country or international commitments.
According to the aims of the investigation, the following three tasks were carried
out:
Task 1—Summary of Solar Thermal Technology Growth.
Task 2—Risk Assessment and Mitigation: This assessment included technological
performance risk, financial/commercial risks, regulatory/institutional risks, and
strategy risks.
Task 3—Market Development Strategy: Following on Tasks 1 and 2, the report
considers the chances of realisation and the bottlenecks of each of the four projects
in the WB/GEF portfolio, including projected market impacts of partial or full
implementation of the portfolio.
As well as completing these tasks, the World Bank also summarised all the CSP
projects currently being considered and developed as well as several institutional,
economic and technical factors that each project faces in its implementation.
36
3 METHODOLOGY
3.1 Outline
To identify systems which are applicable for distributed power generation, energy
and financial modelling was performed to gain an overview of what is available and
how these system configurations can be integrated into urban areas, and more
specifically Wits University. To arrive at a nominal cost of electricity generation, the
affects that these systems would have on Wits University’s electricity usage is also
investigated. The methodology followed is now outlined.
Data Synthesis - Technologies Analysed

Different system configurations and operating performance of the plants in
operation are expected because of the radically different operating conditions. In
order to specify such systems that would be suitable for Wits University, an
extensive comparison of the existing plants was performed.
Data Comparison and Verification

As discussed in the literature review, the Ecostar as well as the Eskom study provide
a very convenient means of comparison between a number of technologies already
in operation. Absolute cost data for each of the reference systems in the studies are
hard to estimate because the systems are all on different levels of maturity.
However the relative distribution of the different cost items is considered to be well
estimated by the approach.
Of the different studies available, the Ecostar study is the most transparent, most of
the cost and performance assumptions are stated explicitly which allows for a very
convenient reference. Methods from the various sources (Section 2.5) are used as
well but it is the Ecostar study that is used throughout this study as a basis for
comparison. The Eskom study is also used but not as extensively because it lacks the
same transparency and is mainly used as a comparison to the Ecostar study.
37
The examination and analysis of the data described above led to the development of
a model comparing the technologies used in the Ecostar Study to that of the Eskom
study. This allowed for rigorous verification of data. The assumptions used in the
various models were also identified, and where applicable criticised or
substantiated. Technical aspects such as plant performance, energy flow and
conversion were also verified. The costing involved in the technologies however was
not verified on an absolute level but just compared to cost data used in other
studies.
Technology Screening and Design Configurations

This model described above was then updated to compare the technologies under
common conditions in South Africa. These conditions are described in Section 3.2.4.
The conclusions from this comparison aided in the technology screening in order to
select systems that will be applicable to urban electric generation. This procedure is
described below.
• Identify several technologies to be used in the comparison.

• Identify functional criteria relevant to distributed urban generation.
• Perform a numerical analysis ranking these criteria.
• Provide a perspective model (tool which ranks the different technologies
according to the desired functions) comparing the chosen technologies.
• Select appropriate alternatives that qualify for distributed urban generation.
A full analysis of these chosen technologies with respect to their installation at Wits
University was then performed, described below.
• Identify appropriate installation sites within the University.

• Analyse the appropriate climate data and its appropriateness for CSP
generation.
38
• Adjust the comparison model for Wits University local data as well as an
appropriate capacity (MW(e)).
• Identify the needs specific to Wits University by analysing its electricity
profiles and usage trends.
Modelling
The conclusions from the above analysis resulted in separate design configurations
being chosen for Wits University. The technical performance of these systems was
then analysed, which included thermal energy flow modelling in Matlab. A model
was then developed that analyses the impact that these technologies will have on
Wits University’s power usage on an hourly basis and how this affects Wits
University’s total bill in order to find a nominal cost of generation.
3.2 Data Synthesis
In order to verify the results obtained by different studies, their methodology

needed to be analysed. The methodology that was followed in the Ecostar study was
briefly touched on in Section 2.5, but the analysis is also common to other studies
and the inputs to these models are explored below.
3.2.1 CSP Technical Performance

Net Annual Solar Electricity
To calculate the net annual electricity generated [kWh] by each of the alternatives
analysed two equations were used, these are detailed below.
Esolar = Aa ⋅ DI a ⋅η s −e (1) (Pitz-Paal et al., 2005)
39
where
Esolar = net annual solar electricity produced
Aa = aperture area of solar field
η s −e = net solar to electric efficiency

DI a = annual direct normal irradiation
η s −e = η Sf ⋅η para ⋅η pbnet ⋅ηrec / pip ⋅η stor (2) ({Pitz-Paal et al., 2005})
where
η Sf = solar field efficiency
η para = efficiency due to parasitics
η pbnet = net power block efficiency
ηrec / pip = receiver/piping efficiency
η stor = storage efficiency
Capacity Factor
The capacity factor is the amount of electricity generated by the plant [kWh]
compared to the rated design capacity.
Solar Capacity factor (S&L, 2003)
Esolar 1
CFsolar = ⋅ (3) ({S&L, 2003})
Wdesign 8760
where
Wdesign = net design output of the power block [kW]
8760 = total amount of hours in a normal year

Esolar = net annual solar electricity produced
40
Net Capacity factor (including hybridisation)
Enet 1
CFnet = ⋅ (4) ({S&L, 2003})
Wdesign 8760
where
Enet = net annual total electricity produced
Solar field Area

Aperture Area
To calculate the total aperture area, Equation (1) above is modified to Equation (5)
provided by Sargent and Lundy which now takes into account the capacity factor.
The goal here is to deliver a certain amount of energy per year according to the
design capacity factor. Another method of calculating the plant area is by designing
a plant according to the peak DNI, but if a plant is designed to give a certain amount
of power at certain times reliability issues result because of the unpredictable solar
resource found in Johannesburg.
Wdesign ⋅ CF ⋅ 8760
Aa = (5) ({S&L, 2003})
η s −e ⋅ DI a
Total Plant Area

According to the results from the Ecostar Study, the cover ratio (GCR), which is the
ratio of solar field area to land area, of all the plants, are found to be between 24-26%
of the total plant area, with the exception of the CLFR plant which makes up 65% of
the total plant area. This ratio is dependent on a number of factors such as the use of
storage, where storage tanks, depending on the capacity, can be quite space
demanding. These ratios are used to calculate the total land area. Other studies
suggest similar ratios. Black and Veatch Corporation (2007) suggest that parabolic
trough systems have a GCR of 30% compared to 70% for CLFR technologies. To
remain consistent, the Ecostar ratios are used in this study.
41
Hybrid Systems
For the purposes of this study, when choosing design alternatives, the added size of
the boiler room, in the hybrid systems, will be ignored. According to the Engineering
Toolbox (2009), a recommended size of a boiler room with a 500kW capacity is 40
m2 which, for the purposes of this study, is considered negligible.
3.2.2 Financial Calculations

Levelised Electricity Cost
“Levelisation” involves calculating a stream of equal cash flows whose Net Present
Value (NPV) is equal to that of a given stream of variable flows. If a project’s
levelised annual cash flow is divided by the annual amount of energy produced, the
result is called the levelised cost of energy. Using a levelised evaluation provides a
simple way to compare alternative projects to each other and is broadly used in the
utility industry.
The levelised electricity cost (LEC) is an indicator of the cost of electricity produced.
This method has been suggested by Sargent and Lundy (2003) as well as the IEA
(1991). CSP plants need to be evaluated on a life-cycle cost basis to determine their
true economic value. This is particularly important for a solar plant since they are
characterised by high initial investment costs that are recovered over a long period
through low operating costs (by virtue of having no fuel expenses).
The total Enet electricity in Equation (6) can be used to verify the electric production
in the Ecostar and Eskom study by using given costs.
fcr ⋅ K invest + K O & M + K fuel

LEC = (6) ({IEA, 1991})
Enet
42
where
fcr = fixed charge rate
K invest = total capital investment
K O & M = annual operating and investment costs
K fuel = annual fuel costs
Enet = annual net electricity.
kd (1 + kd )
n
fcr = + kinsur (7) ({IEA, 1991})

(1 + kd )
n
−1
where
kd = real debt rate
kinsur = annual insurance rate
n = life of plant in years.
Specific Investment
The specific investment is the total cost of the installation (including indirect costs
and contingencies) per installed kW. The total costs of the installation include the
sum of the following components:
• Investment, Solar Field

• Investment, Power Block, Balance of Plant (BOP)
• Investment, Receiver
• Investment, Tower
• Investment, Storage
• Investment, Land.
43
Investment in the receiver and tower refer to the central receiver systems. And the
investment costs for storage are only applied to the relevant systems which
encompass storage.
3.2.3 Model Development and Verification

A Microsoft Excel model has been produced to verify the Eskom and Ecostar data.
The intention of this model is to replicate the models used in the two studies using
the process described in Section 3.2. The model is given in Appendix L. Inputs to this
model include the technical and financial aspects of the plants using the
methodologies followed in Sections 3.2.1 and 3.2.2
The financial assumptions, such as the fcr used in the Ecostar study, have been used.
The intention of this model is to make it fully adjustable so that local conditions such
as insolation, land areas, local currencies etc can be used to reflect results that can
be expected anywhere in the world. The model is later used to compare the
technologies for an initial technology screening.
From this model, it can be concluded that the Ecostar data was verified with the
average error obtained being less than 2 % and this is due to rounding errors from
the data provided in their report. The Eskom data provided was not as detailed as
that given by the Ecostar study. The data verification was therefore not as successful
with errors being between 3-6%.
One major error for the central receiver that uses molten salt in the Ecostar study
was found. For technologies that use only solar energy, these systems will have a
total capacity factor equal to its solar capacity factor. It is only the hybrid
technologies that will have differing capacity factors. When calculating the Enet using
Equation (6) and comparing this to the Esolar calculated using Equation (1), a
difference in the solar and net capacity factors of 15% is found. This implies
hybridisation. Ecostar did not intend their design to include hybrid mode, and fuel
44
costs were hence not accounted for. It was mentioned that the design was adapted
from a plant that includes hybrid operation. The LEC has therefore been adjusted to
take into account the Esolar produced, which will effectively increase the LEC of the
central receiver using molten salt. It is uncertain whether the balance of plant takes
into account the costing of the hybrid components of the plant in the power
block/BOP value. It will be assumed that because the fuel costs were not accounted
for, this value is also neglected.
3.2.4 Model Adjustment

To compare the technologies for use under local conditions the model is adjusted
according to the following criteria, in order to obtain a relevant model. The intention
of these adjustments is to derive a model that will be able to compare the known
technologies under a common local and present day situation. This will assist in the
technology screening.
• Present value: The costing of the technologies analysed in the studies - fuel
prices, maintenance, and hence the LEC - was calculated in foreign currencies
and performed a number of years ago. This value needs to be brought to an
equivalent present day value in South African Rands.
• Local radiation data: The DNI received at different sites will affect the
performance of the plant. If the design capacity factor is to be maintained, the
size of the collector area will change if the DNI changes. Local DNI data
therefore need to be obtained.
• Scaling methods: The data used in the comparison are data for plants that
are usually of quite large scale. An appropriate scaling method is needed in
order to bring these plants to small scale size. This will have an effect on the
costs involved and hence the LEC.
45
Present Value of the Past
The Eskom data provided is based on U.S. dollar prices at the time of writing (2001),
and the Ecostar data is based on Euro prices in 2005. These both need to be
adjusted to find present day values. The future value equation is defined below.
FV = PV (1 + r )n (8) ({Wolf, 1969})
where
FV = the future value of the of the investment
PV = the present value of the investment, often referred to as the principal
r = the interest rate
n = the number of periods for which the investment will be discounted.
The definition of this formula assumes that the present value is the value of today’s
investment and the future value will be the value of the investment n years in the
future compounded at an interest rate r.
For the purposes of this study, the future value FV referred to in Equation (8) is
today’s present value. PV is the actual costs and pricing of the various technologies
in the year of writing. This value needs to be brought forward using the interest rate
r, sometimes called the ‘decay rate’ when finding the ‘present value of the past’
(Wolf, 1969).
Decay Rate
Westney (1997) shows that cost escalations or the variation in prices can be the
result of several factors. General increases in prices can be a result of overall
inflation in the currency; or there can be spot price changes in certain commodities
caused by shortages, built in price changes, or monopolies. These two inflation rates
can be defined as the price inflation and cost inflation respectively. These inflation
46
rates will have an influence on material costs, labour costs, equipment costs, and
other costs such as financing costs.
Chemical Engineering Plant Cost Index

To calculate a suitable decay rate certain indices are available which account for the
price inflation of certain process-equipments. Two indexes, the Marshall and Swift
equipment cost index and the Chemical Engineering Plant cost index are both
recommended (Peters and Timmerhaus, 1991). These two indexes give very similar
results but it is the Chemical Engineering Plant Cost Index (CEPCI) that is used here.
Construction costs for chemical plants form the basis of the CEPCI. This index is
viewed as a better reflection of cost inflation than traditional inflation measures
such as the common consumer price index (CPI). The four major components of the
index are weighted by percentage in the following manner:
• Equipment, machinery and supports 61%

• Erection and installation labour 22%
• Buildings, materials and labour 7%
• Engineering and supervision 10%
The index in based in U.S dollars and even though the Ecostar cost data is published
in Euros, much of their data was sourced in dollars, which allows for an acceptable
adjustment.
Table 3-1: Chemical Engineering Plant Cost Index
Year on year
Year Index
increase %
2000 394.1
2001 394.3 0.05%
2002 395.6 0.33%
2003 402 1.62%
2004 444.2 10.50%
47
2005 468.2 5.40%
2006 499.6 6.71%
2007 525.4 5.16%
2008 609.1 15.93%
Table 3-2: CEPCI Average
Average
2005-2008 8.30%
2001-2008 5.71%
Because the results published by Ecostar were in 2005 Euro, the decay rate to be
used is 8.3%. The rate the Eskom data is adjusted by is 5.71%.
Sargent and Lundy (2003) state the costs decline by a certain percentage with each
doubling of the total number of units produced. These adjusted rates as discussed
do not reflect the effects of the volume of production and learning curve.
Solar Radiation
The effects of a changing input solar radiation will have an effect on the capacity
factor experienced by the plant if the solar field is not scaled correctly. To determine
the size of the adjusted solar field for a change in the direct normal irradiation
(DNI), Equation (1) has been adjusted to form Equation (9). This assumes that the
annual solar-electric efficiency remains unchanged. The resulting solar field size will
be equal to the existing field, multiplied by the ratio of the annual DNI at the existing
site to the annual DNI at the new location.
DI
Aa − new = Aa ⋅ (9)
DI new
The Ecostar study assumes an average annual direct solar radiation value of 2014
kWh/m2a, found in Seville, Spain. The Eskom study uses an annual DNI value of
2900 kWh/m2a. According to the Typical Meteorological Year (TMY2) data obtained
48
from the NREL (1995) for Johannesburg, the average insolation received over the
period of a year is 1780 kWh/m2a which is the value used in the in the analysis. The
details of this data are further described in Section 5.3.
Scaling methods
Sargent and Lundy (2003) state that the economies-of-scale method, for estimating
and evaluating costs of various scaled components, is appropriate. Scaling factors
were used to estimate the cost of a new size or capacity from the known cost for a
different size or capacity. The relationship is based on the following formula:
S 2 Sf
C2 = C1 ( ) (10) ({S&L, 2003})
S1
where
C2 = desired cost of equipment at size (or capacity) of S2
C1 = given cost of equipment at size (or capacity) of S1
Sf = scaling factor
Several of the technologies in the Ecostar study have been scaled up from existing
technologies for comparison reasons. By using Equation (10) it is possible to extract
a scaling factor from the scaling of the technologies. This factor can then be used in
estimating the costs involved with other scaled plants. Where some technologies
were not originally scaled, the average scaling factor was used. This was deemed
appropriate because the scaling factors across the technologies were relatively
constant.
3.3 Technology Screening
In order to select appropriate technologies suitable for urban application, candidate

technologies were identified and by ranking them according to identified functional
criteria, full perspective of practices and operational systems was identified. The
49
methodology followed in the screening of the technologies is further elaborated in
Section 4. The criteria that should be considered for site selection are however given
below.
Site Screening
The selection of an appropriate site for the implementation of a distributed CSP
system in urban areas will be based on certain criteria that will often differ from
that of normal utility plants.
The Bureau of Land Management (BLM) in the USA detailed a report for the National
Energy Policy Implementation Plan, which was to identify and evaluate renewable
energy resources on public lands and any limitations on access to them. The
following is a summary of the criteria identified for CSP systems (BLM, 2003).
Central Generation Technology Criteria:

1. Solar resource is 1700 kWh/m2/a of direct normal radiation (at least) (Stine and
Geyer, 2008)
2. Slope of land area at the site must be less than 5 %, and ideally less than 1 %.
3. Transmission access is within 80 km, and transmission capacity is available.
4. Forty acres is the minimum parcel size.
5. Site must have access to roads or rail within 80 km.
Distributed Generation Technology Criteria:

1. Solar resource is 1700 kWh/m2/a of direct normal radiation.
2. Slope of land area at the site must be less than 10 %.
3. Site must have access to roads.
The following items were also identified but not as the most important screening
criteria.
50
Central Generation Technology Criteria:
• The site must have a low average wind speed (average wind speed < 16 km/hour).
• Water resources must be available.
• The site should be within 40 km of a main natural gas pipeline for some
configurations.
• All vegetation at the site must be removed.
• Federal, state, and local policies are supportive.
• The site must allow structures 5-15 m high. Some technologies could require
structures hundreds of feet high.
• Livestock protection is possible.
• Light reflection at sites near major roads not issues for some technologies
• A population centre should be within 160 km.
Distributed Generation Criteria:

• The site is within 160 km of a population centre.
• Transmission access, water availability, and minimum parcel size are not an issue.
3.4 Application
After analysing Wits University’s load profiles and siting options described in
Sections 5.1 and 5.2, certain design configurations can then be selected which would
include plant capacity, storage integration etc.
To analyse the energy output from the hourly DNI collected over a certain area a
solar field aperture is sized using the average total DNI received over the span of a
year. This aperture area is designed using the method described in Section 3.2.
Using this designed aperture area, the net power output and plant performance are
calculated on an hourly basis. This method is further described below.
51
3.4.1 CSP Plant Performance
Many of the studies performed did not provide full details of their solar-electric
modelling (the conversion of solar energy into electrical energy). In order to verify
their data, the model suggested by Broesamle et al. (2000) has been followed.
The model is made up of two parts, one that simulates the energy balance of the
solar field, being the conversion of solar energy into usable thermal energy. The
second part represents the conversion efficiency of the power cycle.
Figure 3.1 represents the outline of the inputs to the model. It calculates the hourly
thermal power output of the solar field and the electricity yield from the solar direct
normal radiation generated in the meteorology module at a certain location. For the
simulation of the collector field energy output, a simplified stationary model of the
physical properties and behaviour of the collector is applied, as described below.
Figure 3.1: Solar to Electric Efficiency
52
Collector Efficiency
The efficiency of the solar collectors is a function of the geometric efficiency and the
optical efficiency of the mirrors.
ηcol = ξ geo ⋅ηopt (11)
where
ηcol = collector efficiency
ξ geo = geometric efficiency
ηopt = optical efficiency.
Geometric Efficiency
A one-axis tracked parabolic trough collector shows certain losses that depend only
on its geometrical structure. The following geometric losses are considered in the
model:
ξ geo = ξ IAM ⋅ ξ S ⋅ ξ E ⋅ ξcos (12) ({Broesamle et al., 2000})
where
ξ geo = Geometric efficiency
ξ IAM = The incident angle modifier (considers the distortion of the reflected image of
the sun at non-perpendicular incident angles)
ξ S = Shading losses within the solar field
ξ E = Intercept factor - Collector end-losses (the portion of the sunlight that is
reflected outside of the range of the absorber tubes at the end of each collector row)
ξcos = Cosine losses (considers the smaller active area of projection of the collector
due to non-perpendicular irradiation and on the angle of incidence).
53
Optical Efficiency
ηopt = α ⋅ ρ ⋅ γ ⋅τ 1 ⋅τ 2 (13) ({Broesamle et al., 2000})
where
ηopt = optical efficiency
α = coefficient of absorption of the absorber tube

ρ = reflectivity of the mirrors
γ = optical precision of the mirror surface (quality factor)
τ 1 = transmission factor of the mirror glass cover
τ 2 = transmission factor of the glass tube that surrounds the absorber tube
For compound linear Fresnel reflectors this equation is modified to take into
account the effects of the secondary reflector ρ 2 . Equation (13) becomes:
ηopt = α ⋅ ρ1 ⋅ γ ⋅τ 1 ⋅τ 2 ⋅ ρ 2 (14) ({Broesamle et al., 2000})
Thermal Energy conversion

The conversion of the Direct Normal Irradiation into useable thermal energy is
modelled using Equation (15) and Equation (16). The efficiency of the solar field
takes into account the collector efficiency (Equation 11) as well as losses. The
second term in Equation (15) refers to the convection losses in the solar field and
the last term accounts for radiation losses.
 π ⋅U π ⋅ ε ⋅σ 
η Sf = ηcol −
 C ⋅ DI
(TA − Tamb ) −
C ⋅ DI
( TA 4 − Tamb 4 )  (15)

({Broesamle et al., 2000})
54
where
U = convection loss heat transfer coefficient [W/m²K].

C = the factor of concentration of the parabolic trough
TA = mean surface temperature of the absorber tube
Tamb = ambient temperature perceived by the absorber tube exposed to the sunlight
ε = coefficient of emission of the absorber tube surface

σ = Stefan-Boltzmann constant [W/m-²K-4]
DNI = Direct Normal Radiation
Equation (16) takes the total DNI in a certain period and finds the thermal energy
delivered by the solar field.
Qɺ Sf = ASf ⋅ DI ⋅η Sf (16) ({Broesamle et al., 2000})
where
Qɺ Sf = thermal energy delivered by the solar field [kW]
ASf = solar field area
DNI = Direct Normal Radiation
Certain losses in delivering the thermal energy from the solar field to the power
block, Equation (17), result in the net power output (Equation (18)).
Qɺ therm = η piping ⋅η stor ⋅η par ⋅ Qɺ Sf (17) ({Broesamle et al., 2000})
55
where
Qɺ therm = rate of thermal energy input to the power cycle [kW]
η stor = storage efficiency

η piping = piping efficiency
η par = efficiency due to pumping parasitic losses
Qɺ Sf = thermal energy delivered by the solar field
The power generated from the thermal energy generated from the solar field is
calculated using Equation (18).
Pnet = Qɺ therm ⋅η pbnet (18) ({Broesamle et al., 2000})
where
Pnet =net power output [kW]
Qɺ therm = thermal energy input to the power cycle
η pbnet =net efficiency of the power block (including dumping and availability)
Costing
The costing of the design configurations chosen were analysed using average
economic data from multiple sources. The details of this are described in Section
3.2.2.
3.4.2 Energy Modelling

In order to analyse the performance of the chosen design configurations, energy
modelling on an hourly basis was performed using Matlab. The hourly analysis of
the design configurations, when integrated with Wits University’s usage, will result
in certain cost savings that translates into a reduced LEC. The details of the
modelling are further described in Section 6.
56
4 TECHNOLOGY SCREENING
In order to select different technologies considered for application at Wits

University a full screening of the available technologies had to be performed. The
following analysis covers the comparative evaluation of these technologies.
Different selection criteria are identified and the various technologies are evaluated
accordingly.
4.1 Economic Comparison of Existing Technologies
Two major economic comparisons have been performed in the last decade. These, as
previously discussed, are the study performed by Ecostar and one performed by
South Africa’s utility Eskom. Each performed a technology review of similar
operational systems, each with different assumptions.
An Excel spreadsheet model that can be adjusted according to DNI radiation,

capacity and site area has been developed in order to compare the technologies
from the Ecostar and the Eskom study, giving them common assumptions. Although
the results cannot be used as an absolute reference, they are a reasonable base for
the comparison of different reference systems.
To compare the technologies from the two studies, the financial data has been
adjusted to assume the following baseline DNI and plant capacity:
• DNI radiation 2900 kWh/m2a (Upington)

• Plant Capacity 100 MW(e)
These assumptions form the basis for the Eskom comparison and are used in this
model.
57
The economic data in both cases is brought to present value Rands. The plants are
also scaled using the common DNI values as well as the same plant capacity using
the method described in Section 3.2.4.
A summary of the results obtained is provided in Figure 4.1. The solid points in the
figure are technologies represented in the Ecostar study and the hollow data points
represent those adjusted from the Eskom study. For each case, the squares
represent technologies that utilise parabolic troughs as the collector technology; the
circles represent central receiver systems and the triangles represent Dish-Stirling
collectors. The source for the data in Figure 4.1 is shown in the spreadsheet found in
Appendix L.
58
6.05
PT with Thermal Oil
PTwith DSG
CLFR
5.05
CRS Salt
CRS Air
CRS Steam
4.05 CRS Hybrid
Dish stirling
PT Only
PT Hybrid
3.05
PT with Salt Storage
PT with DSG
LEC [Rand/kWhe]
ISCCS
2.05
CLFR Coal
CRS molten salt
CRS Phoebus
1.05 CHIMNEY
Dish-Stirling-1
Dish-Stirling-2
0.05 MTPP
0 20,000 40,000 60,000 80,000 100,000 120,000 140,000 160,000 180,000

Specific Costs [Rand/kWe]
Figure 4.1: Financial Results

59
It can be seen from the graph that the Dish-Stirling concept is by far the most costly
option, both it’s levelised electricity cost (LEC) as well as the specific costs related to
installation are the highest. The details of two separate Dish-Stirling options from
Eskom are provided. The first option (Dish-Stirling-1 in the graph) represents the
cost assumptions used in Eskom’s analysis. These cost estimates are actually
representative of a near term estimate and are not representative of current
installation costs (at the time of writing – Eskom 2001). The second option (Dish-
Stirling-2) shows the current installation costs for 2001, which is viewed as a more
realistic basis for comparison. The Dish-Stirling-2 point also compares closely with
the Ecostar estimate. From this it can be concluded that the Dish-Stirling systems
are the most expensive alternative. Only the Eskom study looked at ISCCS and the
Chimney technologies. The chimney initially appears unfeasible because of the high
costs involved.
These adjusted results are in line with the comparison performed by Eskom and
Ecostar. It is interesting to note that the Eskom data results in a lower LEC
calculation than that calculated by Ecostar. Reasons for this include the fact that
South Africa has specific land costs as well as O&M costs that are lower than those
estimated in Europe.
4.2 Candidate Technologies
The following technologies have been identified because they have either been
tested or put into commercial operation, allowing for performance and financial
data to be collected. It is important to note that these alternatives were chosen
because of the availability of performance data. In most cases they would not be
suitable for small scale distributed power applications. The ranking of these
alternatives is discussed in detail.
60
The majority of the data used here has been taken from the Ecostar study. ISCCS and
the solar chimney technology have been taken from the Eskom study. The MTPP
data has been taken from a study performed by Hassani and Price (2001). A brief
description of these technologies can be found in Appendix B.
1. Standard parabolic trough (SEGS)

2. Parabolic trough with storage (SEGS with storage)
3. Parabolic trough with direct steam generation (SEGS DSG)
4. Compound Linear Fresnel (CLFR)
5. Central receiver with heliostat field and Salt as HTF (CRS - Molten Salt)
6. Central receiver with heliostat field with atmospheric receiver – air as heat
transfer fluid (CRS - atmospheric air)
7. Central receiver with heliostat field with pressurized volumetric receiver
(CRS - Brayton)
8. Integrated Solar Combined Cycle - (Fossil-fired Brayton topping cycle and
solar-assisted Rankine bottoming cycle) (ISCCS)
9. Dish-Stirling engines (Stirling Cycle)
10. Solar Chimney
11. Modular Thermal Power Plant (MTPP)
4.3 Functional Criteria
All functions defined here are intended to be a clear and concise description of what
must be achieved. They have been defined by a verb and a noun, a value
engineering method prescribed by Huber (2008).
Each technology is best suited for different applications. The functions that are most
significant in terms of urban use, and those which can be addressed by different
features from each of the existing technologies, have been identified below.
61
Produce electricity
This is the primary function of the power plant and because all plants (including
fossil fired plants) produce electricity it is important to define this function more
specifically in terms of the capacity factor. The capacity factor is the amount of
electricity produced by the plant [kWh] compared to the rated design capacity.
Capacity Factor
The capacity factor of a power plant is the ratio of the actual output of a power plant
over a period of time and its output if it had operated at full rated capacity the entire
time. The power outages South Africa faces are creating a drop in productivity and
in many cases large economic losses. Energy independence from Eskom is also a
driver of independent power generation. All technologies under consideration can
provide some form of independence from Eskom. This capacity factor will include
technologies under hybrid operation.
Another form of independence, which is not considered here, is the shift from the
reliance on fossil fuels such as oil, gas and coal. As these fuels are depleted, their
demand keeps on rising. Reducing the dependence on these fossil fuels is of high
priority; by emphasizing the reduction of emissions, which is a factor discussed
later, we imply independence from fossil fuels that produce greenhouse gases in
their combustion.
Minimise costs
Cost referred to here is solely the cost of producing electricity. The Levelised Cost of
Electricity (LEC) takes into account initial capital costs, maintenance and operation
and fuel costs. These were therefore not evaluated individually.
Simplify integration
Because urban areas are usually space-constrained, this will be one of the most
important evaluation criteria. The first criterion evaluated is the required floor size
of the solar field and power plant (Power/Area ratio [kW/m2]). The second is the
62
vertical height of the structures. For example building a central receiver on the roof
of an existing building may oppose the aesthetic appeal of the building.
The disjointed nature of the space found in urban areas was also taken into
consideration. One often finds small pieces of land or convenient roof tops that are
too small for the installation of a large scale CSP plant. Hence the modular nature of
some technologies such as micro-steam turbines or the use of organic Rankine
cycles also using smaller turbines could prove to be successful.
Reduce Emissions
This function refers to the reduction of greenhouse gas emissions typically
produced by fossil-fuel power generation facilities. It was deemed less important
because all of the solar technologies reduce emissions to some extent. This criterion
is usually used in evaluating renewable energy in general, as compared to fossil-
fuels.
Different countries and organisations pursue the use of CSP generation for different
reasons. South Africa, as a developing country, even though there is significant
pressure placed on it, is still more concerned with energy security and economic
growth than it is with emission reduction. In some cases the plants’ capacity to
produce electricity, as a function, is a contradiction to this emission reduction
function because a hybrid plant, for example, will score higher in its capacity and at
the same time be penalised with the reduction of emissions.
This function would receive a greater weighting if, for example, the technologies in
comparison included fossil-fuel power plants. A more quantitative judge on the
securing of energy independence from fossil-fuels is the measure of the solar
fraction. The solar fraction is a crude indicator of the different technologies’ ability
to prevent greenhouse gas emissions.
63
Maturity of technology
Here the maturity of the technology refers to its reliability through a record of
demonstrated performance. The reliability of the technology actually refers to
technical risk. Low technical risk is preferred in a project, all else equal. Points were
allocated based on the judgement of the degree of demonstration to date, industry
backing of technology as well as judgement of scaling issues.
Promote Local Industry

Technologies that can be produced using local industries are preferred to those that
need to be imported.
System Safety
Because the technologies will be placed in an urban area, which experiences a lot of
human traffic and thoroughfare, solutions which are least vulnerable to this as well
as to vandalism are credited accordingly. At the same time, the systems cannot pose
a threat to the health and the security of bystanders and maintenance personnel.
4.4 Numerical Analysis
The following numerical analysis is a Value Engineering approach to identify the

most significant functional criteria when evaluating the implementation of CSP
technology in urban environments in Johannesburg. This method has been
recommended by Huber (2008).
In order to prioritize the functions, a numerical analysis is performed using a

‘function numerical evaluation matrix’. This matrix compares functions against one
another to rank their relative importance. The functions are ranked using a score
from 1 to 3, where 1 indicates a minor difference and 3 indicates a major difference.
The scores are then tabulated and from this, the most important functions may be
64
determined. The aim is to create an objective and un-biased method of evaluating
the priority of the functions.
The final outcome of the numerical analysis is to create a cause and effect graph.
This graph separates the most important functions from the less important
functions. The rationale is that addressing the ‘cause’ of the issue will implicitly
address the minor ‘effects’ of the issues, thus focussing on the fewer, critical
functions. The numerical analysis is given below, followed by the cause and effect
results graph.
Figure 4.2: Numerical Evaluation Matrix
Figure 4.3: Cause and Effect Graph
65
As can be noted from Figure 4.2 and Figure 4.3, the four functions that the
technologies will be rated against are:
• Produce Electricity
• Minimise Costs
• Simplify Integration
• Reduce Emissions.
4.5 Perspective Model
Because the performance and costs of small scale power generation modules have
not been well documented, the only resources available are those from technologies
that have been well demonstrated in the past. These technologies range in size from
1 MW for the Dish Stirling concentrators to 80 MW plants for the original SEGS
plants in the Mojave Desert in California. Due to the requirements that distributed
urban generation demands, the selection of the technology that would best be
suited, would therefore not necessarily have the same specifications or capacity as
the existing technologies on the market.
The technique used in the analysis of the alternatives is called perspective modelling
(Huber, 2008). For each of the functional criteria as defined in the first row of Table
4-1 below, each of the alternative technologies is given a score out of ten. The
alternatives score is then multiplied by the functions score (from the numerical
evaluation in Section 4.4), for a weighted score. This process is continued across the
matrix, until alternative one has been evaluated against all the functions. This is
then repeated for all the alternatives. The highest total is the solution that is most
likely to meet al.l the requirements of the evaluation.
66
Accordingly, each of the alternative technologies is ranked from best case scenario
to worst. This is given in the last column in Table 4-1. The results are further
discussed.
Table 4-1: Perspective Model
Functions
3. Simplify integration
1. Produce Electricity
4. Reduce Emissions
2. Minimise costs
Total
Rank
Alternatives
Score 13 11 9 9
1 SEGS 6 78 2 22 4 36 5 45 181 10
2 SEGS with Storage 7 91 5 55 4 36 8 72 254 4
3 SEGS DSG 6 78 3 33 4 36 8 72 219 9
4 CLFR 5 65 6 66 9 81 8 72 284 2
5 CRS Molten Salt 8 104 8 88 2 18 8 72 282 3
6 CRS Atmospheric Air 7 91 4 44 2 18 8 72 225 8
7 CRS Brayton 8 104 9 99 2 18 2 18 239 6
8 ISCC 8 104 10 110 0 0 2 18 232 7
9 Solar Chimney 8 104 0 0 0 0 8 72 176 11
10 Dish Stirling 6 78 1 11 10 90 7 63 242 5
11 MTPP 10 130 7 77 9 81 8 72 360 1
4.6 Alternative Technology Evaluation
Each technology is suited for different applications. These technologies that have
been ranked according to the functions defined in Section 4.3 above will now be
discussed. The ranking of these technologies is by no means ”fair”, for example,
some technologies make use of storage which will increase the capacity factor. By
examining the different features from each of the existing systems, design decisions
and conclusions can be drawn.
67
Produce Electricity
The technologies with the highest capacity factor scored the highest. Even though it
is possible for some of the hybrid technologies to reach the capacity factors of
normal fossil fired power plants, their demonstrated operation is assumed to hold.
With the successful demonstrated operation of a large-scale thermal storage system
at Solar Two (Pitz-Paal et al., 2005), the central receiver with molten salt also scored
high. These systems are expected to be able to be designed for reliable operation
that could extend well beyond early evening peak hours or cloud transients. The
same score was given to the parabolic trough plant with storage which also has
proven dispatchability. Because the atmospheric air central receiver has lower
thermal storage capacity, it scored lower.
The solar chimney does not require direct solar radiation to sustain operation and
can sustain power generation through minor cloud transients. Additionally it
doesn’t require as rigorous start-up procedures. The CLFR technology evaluated
uses direct steam generation which does not have a form of thermal storage and
therefore scored less.
Minimise Costs
The major problem with the integration of renewable energy systems is the price of
electricity produced. The actual cost of the solar field technology is currently too
expensive to compete with utility scale plants. There is significant room for cost
reductions with a reasonable deployment of the technologies in the future. The
technologies were ranked according to their levelised energy cost (LEC). The LEC is
the cost of electricity per kWh and takes into account capital costs, maintenance and
fuel and assumes standard financing costs.
Simplify Integration
Solar Field
When judging the sizing, it is only the size of the solar field that was taken into
consideration. All sets of technologies, for example parabolic troughs, scored the
68
same as they all use the same sized solar fields. The solar chimney scored the least
because of the immense size the solar field demands. The power towers also scored
poorly in terms of space utilisation because they make use of the largest solar field
which in desert areas is not a problem but when it comes to urban areas, the cost of
land or its utility/opportunity cost is significantly higher. CLFR uses the least
amount of land area for the solar field and therefore scored the highest.
Flexibility, Modularity and Practicality

Dish-Stirling and MTPP scored equally high because modularity in urban areas is
critical. In general there is a lack of space and therefore the option of modulating the
technologies is a very attractive option, especially if the plants will be situated on
multiple roof tops. CLFR technologies, even though are not modular, are quite
flexible because they make use of flat, closely spaced mirrors and there is an
opportunity for them to be built on rooftops or used as shading mechanisms such as
above parking lots. Central receivers and the solar chimney scored the least in this
criterion merely because the tower is such a permanent structure and has high
vertical space requirements.
The CLFR technology, because of its compact nature will also be easier to clean. Also
because of its horizontal profile, CLFR technology experiences the least amount of
wind loading and during high wind conditions the mirrors can be adjusted to sit
horizontally and during high hail conditions the mirrors can easily be set to the
vertical. This flexibility allows the CLFR technology to be placed in extreme weather
conditions, often found on the top of high buildings.
4.7 Chosen Alternatives and Discussion
It is noted again that the technologies analysed here, merely serve as the
groundwork for the specific design decisions to be made for distributed urban
generation. This is a general analysis of systems in operation put under comparison
69
in order to gain perspective of what is available. Most of these technologies were
originally designed with different outcomes in mind, for example, to meet peak
loads or evening loads. What is required are distributed generation systems and
what is to be discussed now, is the reasoning behind the scoring, which will lead to
further investigations and initial design decisions.
From the screening analysis performed, it is clear that the Modular Trough Power
Plant (MTPP) scored the highest in the analysis. It is however important to analyse
the reasons for its first place ranking. The MTPP technology scored the highest
points in its ability to produce electricity. The criterion for this function is a high
capacity factor. Here, storage is a design choice. The conceptual design of this plant,
for NREL (Hassani and Price, 2001), assumed 9 hours of storage which contributed
to its high capacity factor. The longest storage for the other technologies in the
comparison was three hours and therefore they did not score as high. It is important
to note here that it is of course possible to design any of the other technologies with
9 hours of storage which would rank them equal to the MTPP system. This is
however a general comparison of existing systems and what is concluded here is
that a system that gives the highest capacity factor is favoured.
Because of the nature and flexibility of the technology, it is possible to select

components from different systems to make up a new concept. The LEC costs will
have to be re-calculated and the accuracy of these calculations (first order) may not
be as credible and hence non-comparable. A detailed life-cycle cost analysis would
need to be performed to find the predicted LEC for the new system.
Purely because of the billing system where the peak demand is billed differently to
the actual usage, hybrid operation or storage will intuitively be the best option at
tackling the need for continued usage during cloud transients. Johannesburg,
especially during the summer, can experience days of cloud cover and in these cases
storage will not solve the problem. Only hybrid operation will work in these cases.
But then the cost of fuel comes into play and whether it will be cost effective
70
operating in this mode through extended periods. If it will be more expensive to
operate in such conditions then the cost of electricity produced through hybrid
operation will need to be compared to the benefits of obtaining energy
independence. It may be the case that the financial disadvantage is too great to
actually install such technology and the primary function will hence remain
unfulfilled. The purpose of installing the technology may then need to be re-
assessed.
In terms of the solar field and collectors, CLFR definitely looks the most promising.
It is by far the cheapest because it doesn’t use parabolic shaped mirrors and
operates with one-axis tracking. This, however, does decrease the efficiency of the
plant.
The use of an ORC is also very advantageous because of its low operating
temperatures. Again the efficiency of these cycles is low and if used in conjunction
with CLFR collectors, the efficiency of the entire cycle will be very low and hence
may demand greater space requirements and perhaps nullify the space advantages
of the CLFR collectors. These differences are investigated further.
Technologies that are clearly not feasible for small scale generation are listed below,
with reasons.
Central receivers: The central receiver with molten salt ranked second mainly
because of its ability to produce electricity at very low costs. They have the potential
for large scale integration but require the amount of land space for the solar field, as
well as vertical space for the tower. They are therefore not suited for modular
production. Because the Brayton cycle plant requires high inlet temperatures (only
reachable in CRS plants), it is also eliminated.
71
Solar Chimney: For the same reasons that central receivers are unsuitable for small
scale distributed generation, solar chimneys are also unsuitable. They require vast
amounts of space and also produce electricity at high costs.
Dish Stirling: The dish Stirling concept, on all other accounts, may seem perfect for
modular integration but they currently produce electricity at very high costs and are
therefore eliminated.
DSG: Direct Steam generation is an advanced technology and also has safety
concerns and will thus not be further investigated.
The technologies best suited are the MTPP and CLFR concepts. These concepts are
investigated further, with and without storage and hybridisation. These are also
evaluated with two different power cycles; normal steam Rankine Cycle and an
Organic Rankine cycle.
72
5 APPLICATION
As noted, the University of the Witwatersrand is used as a case study for the
possible installation of a pilot scale CSP generation system. The following section
analyses the feasibility of such CSP systems.
5.1 Wits Electricity Profiles
Wits University is separated into six separate campuses which are all billed
separately. These campuses include: Gate House (East Campus), Raikes Road (West
Campus), Wits Business School, Wits Education Campus, Wits Medical School,
University Corner. Wits East and West Campus consume the most electricity. West
Campus’ electricity profiles are examined here. All data is taken from the Metoring
Online site that controls bills for City Power customers in Johannesburg (MOL,
2008).
The monthly bills are calculated on a per kWh basis as well as charging for the peak
kVA utilised. This peak kVA charge is quite significant, as shown in a typical monthly
bill, in Appendix G.
The CSP systems will be used to try to decrease the daily peak load because it is this
that is responsible for the Maximum Demand charge. The Maximum Demand
experienced is different on all campuses and occurs between the hours of 7 am to 8
pm, with the maximum peak experienced at mid-day. Solar technology is therefore
an obvious solution because the sun shines during the day. On an annual basis, this
full load scheme represents a 54.1% (=13/24) capacity factor.
Also seen in Appendix G are two electricity profiles, the first being that for the
month of November 2007 and the second for June 2008. Each profile indicates the
distribution of usage in a twenty four hour period, for a typical summer and winter
73
month during University term. Both show a common peak at 12h00 and again at
19h00. Because the peak kVA contributes a significant amount to the monthly bill
(approximately 1/3), it would be ideal to decrease these two peak levels. For
summer usage, this means bringing the usage down from a peak of 2100 kW to 1600
kW and in winter from 2700 kW to 2200 kW (production from 07h00 to 20h00).
This is the equivalent generation capacity of 500 kW.
5.2 Site
Actual site selection for the installation of the CSP system is beyond the scope of this
study. However general application and sizing of the system is important in
determining the feasibility. The site plans for Wits University Main Campus were
obtained and some of the building areas were simply measured. A list of potential
sites is given in Table 5-1. These figures represent the roof areas of the various
buildings or open areas and serve merely as a guide to conceptualise the possibility
of system integration.
Table 5-1: Potential Site Areas
West Campus [m2]

Parking Lot Outside Hall 29 6365
DJ du Plessis Building 9900
New Commerce 2220
Commerce Library 2400
West Campus South Side Parking 8000
Genmin Laboratory 1750
East Campus [m2]

Old Mutual Sports Hall 2400
Senate House 7600
Hillman Block 1820
NWE Building 2625
74
SWE Building 2450
Only larger buildings with symmetrical rectangular shapes are given here. These
sizes are not entirely accurate because only the main rectangular sections were
measured. The suitability of the roofs, in terms of structural, practical and aesthetic
factors is beyond the scope of this report.
5.3 DNI Data
The weather data used in the analysis is Energy Plus Weather (EPW) format in SI
units. The format is simple text based data based on the TMY2 (typical
meteorological year) format but rearranged to facilitate visual inspection of the data
(NREL, 1995).
A TMY provides a standard for hourly data for solar radiation and other
meteorological elements that permit performance comparisons of system types and
configurations for one or more locations. It represents conditions judged to be
typical over a long period of time. Specifically, for Johannesburg, the data represents
a typical meteorological year from data collected over 30 years from 1961 to 1990.
The generic site chosen in this case is Wits University. Due to the lack of solar data
specific to Wits, generic data for Johannesburg (Latitude -26.13, Longtitude 28.23
and elevation of 1700m) which is available from the NREL website (NREL, 1995), is
used in the analysis.
Seville, Spain, has a typical DNI of 2014 kWh/m2a (Pitz-Paal et al., 2005) and has
been the site for many CSP applications. Johannesburg’s DNI is typically 1781
kWh/m2a (shown in Appendix C), a value lower than that of Seville but above the
recommended 1700 kWh/m2a (Stine and Geyer, 2008).
75
The TMY2 data of 1781 kWh/m2a is typically less than that found using Figure 2.2
(1825-2190 kWh/m2/year), but resolution of the data obtained from this DNI map
is lower than the TMY2 data which is why the TMY2 data is used. Appendix C
provides details of this data. Even though the summer months in Johannesburg yield
a higher peak DNI, it is actually the winter months that provide a more consistent
average. This is due to the high amount of cloud cover experienced in summer.
5.4 Design Configurations
In order to assess the potential of CSP generating facilities at Wits University, the
following combination of technologies have been chosen for the assessment. These
alternatives have been chosen following from the technology selection criteria
discussed and selection detailed in Section 4.7.
1. Parabolic Trough with normal Steam Cycle, no storage, no hybridisation

2. Parabolic Trough with normal Steam Cycle, with storage
3. Parabolic Trough with normal Steam Cycle, with hybridisation
4. Parabolic Trough with Organic Rankine Cycle, no storage, no hybridisation

5. Parabolic Trough with Organic Rankine Cycle, with storage
6. Parabolic Trough with Organic Rankine Cycle, with hybridisation
7. CLFR with normal Steam Cycle, no storage, no hybridisation

8. CLFR with normal Steam Cycle, with storage
9. CLFR with normal Steam Cycle, with hybridisation
10. CLFR with Organic Rankine Cycle, no storage, no hybridisation

11. CLFR with Organic Rankine Cycle, with storage
12. CLFR with Organic Rankine Cycle, with hybridisation
76
5.4.1 Output Capacity
In order to compare the different options a standard electric output of 120 kW(e)
has been chosen. This size was determined by considering different applications
currently in operation. As an example, Freepower, being one of the many ORC
technology suppliers, provides 120 kW(e) Organic Rankine Cycle Turbine
Generators which are closed cycle electrical power generation systems driven by an
external heat source with no internal combustion being needed. This or a similar
generator would be used in the implementation. Technical aspects of the Freepower
generator can be found on their website (Freepower, 2008).
The initial site sizing which would integrate a parabolic trough system to provide
the required thermal energy input to such a steam turbine generator is
approximately 4000 m2 and 1400 m2 for a CLFR collector field. (This was calculated
using the initial model developed in Section 4.1 using the Ecostar specifications).
Because of the need for modularity, this 120 kW(e) system is used as the reference
size. In order to make the required impact on Wits University’s electric bill, and to
create a system that will provide better energy security, a bigger capacity will need
to be installed. The plants are to be modularly integrated and will hence be installed
in multiples of the 120 kW(e) reference plant. By examining Wits University’s usage
profile for West Campus in Section 5.1, it is concluded that 500kW of production will
satisfactorily decrease the peak day time usage without demanding high production
in hours with no sunlight. Therefore, 480 kW(e) is chosen because this is a multiple
of the 120 kW(e) reference plant. These systems, depending on space availability,
may be located at the same site or, with modularity in mind, at separate sites.
5.4.2 Reference Plant

The alternative selections with no storage or hybridisation will be designed to have
an average capacity factor of 20%. Assuming constant solar load, this represents
solar electric production for 5 hours per day which will be sufficient to meet mid-
77
day peak loads. Because of weather effects, some days will experience no sunshine
and therefore is compensated by days with considerable sunshine, generating
electricity for more than the 5 hour peak period.
5.4.3 Storage
The storage alternatives will be designed to give an average capacity factor of 30%
which make will provide approximately 3 hours of storage at the design input. The
solar multiple for a plant will be 1.5 (=0.3/0.2). The actual size of the plant will be
greater than 1.5 times the size of the reference plant because of the effect of the
storage efficiency of these alternatives.
The design choice for solar storage is a one tank system because of the space
constraints at Wits. It is suggested that the storage system, if implemented, is an
active storage system based on the thermocline design (see Appendix A). The
storage efficiency for such a system is unknown and beyond the scope of this report.
For the purposes of this study, the efficiency of the storage system is assumed to be
94.7% which is the same storage efficiency found in the Ecostar study. This
efficiency is applied to the entire system, which in effect will tend to under-estimate
the net solar to electric efficiency. This is satisfactory in this case but when a full
analysis is performed using Matlab, this efficiency is only applied to the thermal
energy that enters the storage system.
5.4.4 Hybridisation
As discussed, the electricity usage at Wits University is significantly higher than the
base load between 7am and 8pm giving a capacity factor of 54.1%. By making use of
hybridisation significant LEC production savings can be achieved. This will also have
a significant effect on bringing down the peak demand experienced by Wits
University, in turn bringing down the peak price paid for electricity.
78
5.5 CSP Plant Performance
The model provided by the Ecostar study does not provide full details of its energy
modelling; the model suggested by Broesamle et al. (2000) has therefore been
followed. Only the data relevant to the chosen alternatives will be analysed.
5.5.1 Parabolic Trough

The Collector efficiency of the parabolic troughs is 67%. This is calculated using
Equation (11) with a geometric efficiency of 89% (Broesamle et al., 2000) and an
optical efficiency of 76%. This is based on the design data of the LUZ-2 reflectors.
The physical parameters representing the LS-2 parabolic trough collectors which
were used in the SEGS plants in California are listed below.
Table 5-2: Optical Characteristics of the Parabolic Trough System
ρ1 =0.93
τ 1 =0.98
τ 2 =0.95
α =0.94
γ =0.93
Solar field efficiency

The efficiency of the solar field ( η Sf ) given by Equation (15), is a function of the
convection and radiation losses and is equal to 51.23%. The details of the
convection and radiation losses are described below.
Convection losses described by the second term in Equation (15)

π ⋅U
[ (TA − Tamb ) ] are equal to 3.5%. Data used in this calculation are provided in
C ⋅ DI
79
Table 5-3. This data is provided by Broesamle et al. (2000) and as stated represents
the LUZ-2 type reflectors. The DNI data is taken locally for Johannesburg and is not
the design point radiation value (this value will be taken as the peak value
experienced in the year). It is hard to say to what extent the use of local radiation
data will affect the temperatures experienced by the absorber. For ease of
calculation it is assumed here that this effect is minimal and is not accounted for.
Table 5-3: Convection Losses for Parabolic Trough Collectors
DNI Direct Normal Radiation 800 W/m2

U Convection Loss Factor 2 W/m2K
C Factor of Concentration 72
Ta Mean Surface Temp of Absorber Tube 653 K
Tamb Ambient Temp 330 K
Convection losses 0.035
Radiation Losses are accounted for in Table 5-4 and are equal to 12.6%, using
Equation (15). Again the data used is taken from Broesamle et al. (2000).
Table 5-4: Radiation losses for Parabolic Trough Collectors
sigma Stefan-Boltzmann constant 5.6704E-08 W.m-2K-4

e coefficient of emission 0.24
Radiation losses 0.126
80
5.5.2 CLFR
The optical efficiency of the CLFR system will be slightly different from that of the
parabolic trough because the Fresnel receivers make use of a secondary reflector
which in turn decreases the efficiency of the system. The data used to calculate the
optical efficiency has been taken from the Solarmundo study (see Table 5-5) where
a 2500 m² prototype collector system was built and tested (Häberle et al., 2002).
The extra variable listed is the secondary receiver efficiency ρ 2 . The optical
efficiency is therefore 68% which is a significant drop from that found in the
parabolic trough collectors. (This efficiency takes into account the surface quality
factor of 93%). This efficiency has a high correlation to that suggested by Mills et al.
(2003) who claim an array could collect a high 75% of the available beam if used
with a reflector with ρ1 = 91%, and 66% with an inexpensive glass reflector.
Mills and Morrison (2000) describe the design of a CLFR power plant and the effect
of the different configurations of the designed plant. These configurations such as
north-south orientation, longitude, incline angle of the slope etc all have an
important effect on the performance of the plant. A geometric efficiency of 80% is
assumed which results in a collector efficiency of 54.5%.
Table 5-5: Optical Characteristics of the Linear Fresnel System
ρ1 =0.91
τ 1 =0.95
τ 2 =0.95
α =0.94
γ =0.93
ρ 2 =0.95
81
Solar Field Efficiency
The efficiency of the solar field η Sf is calculated using the same method described by
Broesamle et al. (2000) for the thermal modelling of parabolic troughs. The
efficiency of the solar field is 40.1% which takes into account the convection and
radiation losses as accounted for below. Because of the lower operating
temperatures the losses are lower than that of the parabolic trough. The
concentration factor is half of that of the parabolic troughs. These values including
the operating temperatures have been suggested by Mills et al. (2003). The DNI
value is Johannesburg’s local value as previously discussed.
Table 5-6: Convection Losses for CLFR Collectors

Convection losses 0.059
Table 5-7: Radiation Losses for CLFR Collectors
sigma Stefan-Boltzmann constant 5.6704E-08 W.m-2K-4

e coefficient of emission 0.12
Radiation losses 0.085
5.5.3 Thermal Energy Flow

The plants are designed for a capacity factor of 20% and 30% as discussed. This
represents a net electric output calculated using Equation (3) where all the
82
electricity will be produced from thermal energy provided from the solar field. The
alternatives are analysed using two different power cycles; a steam cycle and an
ORC. The power block design specifications are given in Table 5-8.
Because of the lower efficiency of the ORC, the size of the field will be larger in order
to collect the required amount of thermal energy to produce the same amount of
electricity.
Table 5-8: Power Block Design Specifications
Design Output 120 kW(e)

Efficiency (npb) Steam Cycle 0.355
Efficiency (npb) ORC 0.23
The required thermal energy delivered by the solar field Qɺ therm is determined from
Equation (18) where the values are initially calculated on an annual basis. This
value for the thermal energy input assumes that the power cycle efficiency includes
that of pumping losses. The solar field efficiency η Sf is calculated using Equation
(17). The values used in this equation have been taken from the Ecostar study
where the efficiencies are given in Table 5-9 and the values calculated in Section
5.5.1 and 5.5.2.
Table 5-9: Parabolic Trough Efficiencies
Parasitic losses 0.908

Storage 0.947
Piping 0.851
Summary
This method has been verified against the results obtained in the Ecostar study and
details of this are given in Appendix D. The aperture area of the systems is
calculated using Equation (16) (results given in Section 7.1). The results for the
alternative plant configurations are shown in Table 5-10.
83
Table 5-10: Design Efficiencies
Solar Power Block Solar to

Power Optical Geometric Collector Convection Radiation Parasitic Piping Storage
# Collector Type Field Efficiency Electric
Cycle Efficiency Efficiency Efficiency Losses Losses Losses Losses Efficiency
Efficiency [net] Efficiency
1&3 Parabolic Trough Steam Cycle 75.7% 89.0% 67.4% 3.5% 12.6% 51.2% 90.8% 85.1% 100.0% 35.5% 14.1%
2 Parabolic Trough Steam Cycle 75.7% 89.0% 67.4% 3.5% 12.6% 51.2% 90.8% 85.1% 94.7% 35.5% 13.3%
4&6 Parabolic Trough ORC 75.7% 89.0% 67.4% 3.5% 12.6% 51.2% 90.8% 85.1% 100.0% 23.0% 9.1%
5 Parabolic Trough ORC 75.7% 89.0% 67.4% 3.5% 12.6% 51.2% 90.8% 85.1% 94.7% 23.0% 8.6%
7&9 CLFR Steam Cycle 68.2% 80.0% 54.6% 5.9% 8.5% 40.1% 90.8% 85.1% 100.0% 35.5% 11.0%
8 CLFR Steam Cycle 68.2% 80.0% 54.6% 5.9% 8.5% 40.1% 90.8% 85.1% 94.7% 35.5% 10.4%
10&12 CLFR ORC 68.2% 80.0% 54.6% 5.9% 8.5% 40.1% 90.8% 85.1% 100.0% 23.0% 7.1%
11 CLFR ORC 68.2% 80.0% 54.6% 5.9% 8.5% 40.1% 90.8% 85.1% 94.7% 23.0% 6.8%
84
5.6 Income and Expenses
5.6.1 Specific Costs

Determining the actual LEC of the CSP design alternatives here is merely a first order
estimate that follows the method prescribed by the IEA (1991) given in Section 3.2. To find
the relevant LECs the total costs of the system need to be determined. To find the total
costs, the calculation has been simplified to incorporate only the major costs of the
systems. Only data relevant to the alternatives being analysed has been used.
These costs are taken strictly as specific costs which is the price in Euros (2008 Euro/kW
or Euro/m2 solar field –where applicable). The total specific costs are converted to Rands.
The following exchange rates were used in the conversions (Taken on 18 November 2008-
(XE, 2008)):
Table 5-11: Exchange Rates
Euro/dollar 0.78
Rand/Euro 13.09
Data from multiple sources have been utilised. A summary of the findings is given below
and the results are given in Appendix H.
Ecostar
The model described in Section 3.2.3 has been further adjusted to take into account the DNI
radiation found in Johannesburg (1780 kWh/m2a) as well as the net output of 120 kW for
the alternatives. These costs have been adjusted from 2005 Euros into 2008 Euros.
Eskom
The same procedure has been performed on the Eskom data, but here the prices are
adjusted from 1999 U.S. dollars into 2008 Euros.
85
MTPP
The report described in Section 2.5 on the MTPP systems has also been incorporated into
the comparison. The 2001 dollar costing has been adjusted to present value Euros.
BIO ORC
Obernberger et al. (2002) describe the specific costing of the installation of a biomass fired
ORC plant in Austria. It is only the costing of the power block that has been taken here in
order to compare these costs to that in the MTPP study. Again the present value of the 2002
Euro quote is found.
Solarmundo
The data described in Section 2.5 has been used in the data comparison. Solarmundo
performed a comparison study between a 50 MW parabolic trough plant as well as a 50
MW CLFR plant.
5.6.2 Levelised Electricity Cost

It is important to note that, in evaluating the attractiveness of the CSP technology, the LEC
cannot be compared to the current costs of electricity as supplied by City Power. LEC
calculations are very sensitive to the economic assumptions used, which do not match the
actual accounting procedures used to set electricity tariffs. In addition, an assumption of
equivalent service is implicit in technology comparisons using an LEC approach; the only
thing that varies between the options is the cost of production. Because of this it makes it
inappropriate to compare the cost of energy to that of a coal fired power plant for example.
Here the plants will vary in terms of the operating characteristics, dispatchability,
modularity and environmental benefits. The results of the LEC analysis will therefore only
give a general indication of the economic value to Wits University and more importantly
serve as a good relative comparison.
Equation (6) has been used to determine the LEC of the selected systems. To calculate the
LEC four components need to be determined.
86
• Investment costs
• Fuel Costs
• Operating and Maintenance costs
• Total Electricity Production.
Investment costs
The levelised required revenue is a fixed charge given by the first term in Equation (6). It
comprises two components-the total investment and the fixed charge rate.
The total investment here is found by taking the average specific investment costs from
each of the above mentioned sources for a capacity of 120 kW. The investment into the
solar field and the actual land is based on the aperture area required to deliver enough
thermal energy to satisfy the required capacity factor for each of the technologies. The
results of the analysis are shown in Table 5-13.
Land costs being significantly less in South Africa than in Europe contributed to Eskom’s
claim that they would be able to produce the cheapest solar electricity in the world (van
Heerden, 2001). Because this study looks at the possibility of a distributed plant in urban
areas, land will be less available and therefore will come at a greater cost. However, this
may not always be the case. If a private firm for example already owns the space which has
no other utility value then this land will, in effect, be free. For Wits University’s case the
land costs are assumed to be zero. The opportunity costs involved with the loss of land are
ignored.
Fixed Charge Rate

Because the installation of this system would not be a financial investment decision, the use
of debt to finance the project would be inadvisable. The capital costs used for the
installation would have an opportunity cost that would be equal to the loss in interest
earned in alternative investments. It is assumed here that the opportunity cost will be 10%
87
which represents a typical return from Standard Bank’s money market in South Africa (SB,
2008).
The FCR assumptions are given in Table 5-12 and the FCR is calculated using Equation (7).
The insurance rate used in the Eskom study was 0.5%. A value of 1% was chosen because
the system alternatives chosen have very little operational experience.
Table 5-12: Economic Assumptions
Annual Insurance Rate 1.00%

Interest Rate 10.00%
Depreciation Life 30.00
FCR (Fixed Charge Rate) 11.61%
Fuel Costs
To determine the effects of hybrid operation on the cost and production, a fuel source
needs to be selected. Egoli Gas provides Johannesburg with natural gas that is a convenient,
easily tapped and cost effective energy source (EG, 2009). Egoli Gas' pipeline network
ensures that gas is instantly available at point-of-use, and they also guarantee that delivery
delays will never occur. The gas is lighter than the air and disperses easily and harmlessly
into the atmosphere, making it relatively safe. For the above reasons natural gas was
chosen as the fuel source for hybrid operation. A detailed analysis of different fuel sources
is beyond the scope of this report. The fuel specifications and tariffs are provided in
Appendix F.
Operating and maintenance costs

The operating and maintenance costs for the CSP systems are shown in Table 5-13. The
data used to find these average values is given in Appendix H.
88
Table 5-13: Average Economic Data
Economic Data
Source Average
CSP System Parabolic Trough CLFR Steam Cycle 0RC
Capacity 120kW 120kW 120kW 120kW
Specific Investment cost for Solar Field (*/m2) R 4,159 R 2,009
Specific Investment cost for Power Block [*/kW(e)] R 15,845 R 14,536
Specific Land Cost [*/m2] R0 R0 R0 R0
Specific Investment in Storage[*/kWh(th)] R 647 R 647 R 647 R 647
Surcharge for Construction, Engineering and Contingencies % 20.00% 20.00% 20.00% 20.00%
Average O&M costs [% of total capital costs] 2.88% 3.75% 2.00%
5.6.3 Income Sources

There are three possible income streams for renewable energy electricity generators.
These are selling physical electrical power through a Power Purchase Agreement (PPA)
into the electrical grid at prevailing electricity (energy) market price, Certified Emission
Reductions (CERs) trading through the Clean Development Mechanism (CDM) of the Kyoto
Protocol and issuing of Tradable Renewable Energy Certificates (TRECs) (Schaffler, 2007).
These are further discussed below.
Power Purchase Agreement (PPA)

South Africa has a high level of renewable energy potential and presently has in place targets of
10,000 GWh of renewable energy by 2013 (DME, 2003). To contribute towards this target and
stimulate the renewable energy industry in South Africa, there is a need to establish an
appropriate market mechanism. It is currently possible to sign a PPA with Eskom who is
designated as the single buyer in South Africa (Eskom, 2009), however to stimulate renewable
sources of energy NERSA is currently working on he implementation of a Renewable Energy
Feed-in Tariff (REFIT) for South Africa (NERSA, 2008). The Renewable Energy Purchasing
Agency (REPA), which will act as the single buyer for renewablevenergy dispatched from the
independent power producers (IPPs), is still under the authority of Eskom. So whether licenses
are correctly handled and the playing field is actually level, is questionable. NERSA have
89
indicated that these ‘loose ends’ would be dealt with over the next three months, during which
time the precise tariff flow-through arrangements for cost recovery would also be finalised.
Feed-in Tariffs are, in essence, guaranteed prices for electricity supply rather than
conventional consumer tariffs. The basic economic principle underpinning the REFIT is the
establishment of a tariff that covers the cost of generation plus a "reasonable profit" to
induce developers to invest. This is quite similar to the concept of cost recovery used in
utility rate regulation based on the costs of capital. Under this approach it becomes
economically appropriate to award different tariffs for different technologies. The PPA
signed with at the appropriate REFIT should also be certain and have a long term
guarantee to allow for project financing to be raised by the project.
Table 5-14 below is a comparison between the REFITs that are currently being offered overseas.
The max capacity specifies the maximum size of a permissible plant that will qualify for the
tariff. Other tariffs for larger plants are available. The terms of the Power Purchase Agreement
(PPA) for the REFITs may include a certain minimum capacity or designs that include a set
storage or hybrid capacity. For example Spain’s REFIT dictates that only 15% of production is
allowed from hybrid operation (Geyer, 2007).
Because South Africa is a developing country we will be able to make use of lower labour
and with our abundance of land, land is also cheaper, which both translates into a lower
LEC. It is hard to estimate what the CSP REFIT will be for South Africa but this value, in
order to attract international investors, will have to be competitive with REFITs found
elsewhere. The lowest value of the REFITs (R2.05/kWh) in Table 5-14 will be used for the
analysis in this study. It is also assumed that this value remains constant over the life of the
plant. It is also assumed that this value will have no restrictions on the technology used.
90
Table 5-14: International CSP REFITs (Geyer, 2007)
Country Max Capacity REFIT [/kWh] REFIT [/kWh]

France 12MW € 0.30 R 3.93
Germany - € 0.46 R 6.02
Greece 5MW € 0.24 R 3.14
Israel 20MW $ 0.16 R 2.05
Portugal 10MW € 0.21 R 2.75
Spain 50MW € 0.27 R 3.54
The main motivator for the use of CSP technology at Wits University is energy security. The
main purpose of a REFIT is to stimulate the market, allowing IPPs to invest in renewable
energy technologies that, without a guaranteed income, would otherwise be unfeasible. If
Wits University (or any other entity) is interested in an alternative source of energy to fulfil
energy security and supply issues, it means that they would be generating electricity for
themselves and obviously would not be able to sell the electricity to the designated
authority under the REFIT. However, it has been decided to investigate generating costs of
electricity which would be sold under the REFIT.
Clean Development Mechanism (CDM)

The Clean Development Mechanism (CDM) is an arrangement under the Kyoto Protocol
allowing industrialised countries with a greenhouse gas reduction commitment (called
Annex 1 countries) to invest in projects that reduce emissions in developing countries as an
alternative to more expensive emission reductions in their own countries. It has been
operational since 2006 and had registered more than 1000 projects equivalent to more
than 2.7 billion tonnes of C02 reduction (CDM UNFCC, 2009).
A crucial feature of an approved CDM carbon project is that it has established that the
planned reductions would not occur without the additional incentive provided by emission
reductions credits, a concept known as "additionality" (UNFCC, 1998).
91
Finance from the CDM
Certified Emission Reductions (CER) or “Carbon Credits” can be sold at any stage of the
development or implementation of a CDM project. CERs are traded on an internationally
regulated market at a price per tonne of carbon dioxide reduced. Over the past few years
this value has ranged between €5-€20/ ton CO2 equivalent.
Because the proposed development will involve the replacement of electricity by a

renewable source of energy, the baseline calculation used to analyse the CER income is
performed using data from Eskom’s current generation mix and the emissions associated
with the generation output. The data used has been taken from Eskom’s annual report. The
figure used represents the Eskom average CO2 figure. Eskom have calculated the carbon
emission factor to be 1,2kg/kWh in accordance with the CDM approved consolidated
methodology 0002. Further information can be obtained in Eskom’s annual report (Eskom,
2008).
It is uncertain what policies will change ‘post Kyoto’, however it is recommended that any
CSP project is registered as a CDM project as there are various financing options that are
attractive such as the forward selling of CERs that can provide development finance. These
various options, because of their uncertainty and variety have been excluded from analysis
in this study and it is recommended that they be incorporated into a full financial analysis.
Tradable Renewable Energy Certificates (TRECS)

The concept of TRECs is based on separating the various attributes of renewable resource-
based energy provision from the physical energy carrier, electrical or otherwise. TRECs
represent all of the benefits (“green” attributes, excluding greenhouse gas mitigation)
associated with the generation of electricity from renewable energy resources. A major
advantage, apart from the “extra” income stream, is that TRECs can be traded worldwide
and separately from the electricity grid infrastructure, thereby avoiding the complexities of
use-of-grid system charges or grid access problems. TRECs are only applicable to
renewable energy and can be issued and traded for all types of renewable energy including
92
non-electrical renewable energy systems, such as solar water heating systems, which
would offset fossil-based electricity production requirements (Schaffler, 2007).
In March 2008 the South African National Tradable Renewable Energy Certificate Team
(SANTRECT) was formed by the DME with an aim to facilitate and coordinate the
establishment of Issuing Body (IB) as Non-Profit Organisation (NPO) that will be
responsible for registering, issuing, transfer and redeem certificates in South Africa. The
SANTRECT is in the process of developing the constitution and the IB will be registered
thereafter. The SANTRECT is intending to register an IB, and hence establishment of IB
NPO by March 2009 (DME, 2009).
The use of TRECS to finance the project is certainly an option and will allow for energy
security by not having to sell away the electricity under the REFIT. However, there is no
solid framework in place and has been excluded from the analysis. It is recommended that
this be investigated when in place.
93
6 ENERGY MODELLING
In order to determine a nominal cost of electricity for the chosen alternatives

thermal modelling as well as the analysis of the West Campus bill has been
performed using Matlab (© 2005). This was done for all the reference plants, as well
as for the 480 kW(e) chosen capacity. The models and details are outlined below,
and the results are found in Section 7.
• DNI synthesis
• Design Analysis and Thermal Modelling
• System integration and bill calculation
6.1 DNI Synthesis
The hourly DNI data obtained in the TMY2 format comes in the form of a CSV
(comma separated values) which is merely a list of 8760 values (number of hours in
year). A short code was written that assigns a month, day and hour to each value,
allowing for data analysis.
6.2 Design Analysis and Thermal Modelling
The hourly electric performance of the of the twelve design variations described in
Section 5.4 was then modelled. This takes into account the design parameters
described in Section 5.5 and the DNI data for Johannesburg, described above. These
input parameters are shown in Table 5-10, Table 7-1 and Table 7-2 (see Section
7.1). The model can be adapted to find the performance of these systems for any
design requirement as well as any DNI input.
94
The hourly thermal energy collected by the solar field ( Qɺ Sf ) is found using the input
DNI data in Equation (16). It is merely the DNI [W/m2] multiplied by the aperture
area of the field [m2] to find a value for the initial thermal energy collected [kW].
The thermal energy that is delivered by the system to the power block ( Qɺ therm ) is
calculated using Equation (17). To overcome the thermal inertia required by the
solar field and power block (especially on start-up), the minimum design thermal
energy delivered to the power system is taken as 25% of the design thermal input. If
the system incorporates no storage, the model assumes that this energy is dumped.
In actual operation, this thermal energy (radiation values usually below 200 W/m2
(Broesamle et al., 2000)) would be used to ‘warm up’ the system. Anything over the
maximum design thermal load will be dumped. In actual operation instances when
the DNI radiation received is higher than the design point radiation, part of the field
will be set ‘off concentrate’.
The electric energy delivered by the system takes into account the efficiency of the
net solar-to-electric efficiency uses the gross efficiency of the power block in
Equation (1).
Storage
For the systems that incorporate storage, any thermal energy over the maximum or
under the minimum design load is summed as the thermal energy for the day. This
thermal energy is then converted to electric energy and dispatched at 18h00 every
day. If this thermal energy is more than the required energy to provide an hour’s
worth of design electric output then the remainder will be used in the second hour
(19h00), this is again repeated for the third hour (20h00).
Hybrid Operation
Systems under hybrid operation will run under full rated capacity from 07h00 till
20h00 which is 13 hours per day or a capacity of 54.1%. It’s solar-only capacity
factor will be the same as the system with no storage (20%). The Matlab model finds
95
the difference between the rated capacity (120kW) and the electric output from the
solar system and adds that difference as a hybrid input. This keeps the power
generation constant.
The model then finds the average monthly thermal flow from the input radiation to
the output electric generation. It also finds the average electricity generated using
thermal storage and under hybrid operation. The average thermal energy flow for
each hour of the day in one year is also calculated.
Three Excel files are then generated which were used as the input to calculate the
exact effect that the systems have on the usage and demand of Wits University.
These files include the solar-electric generation of each system, as well as the
storage and hybrid outputs for the relevant systems. This is described below.
6.3 System Integration and Bill Calculation
A Matlab code was written by Brink (2008) in order to test the effect that the new
electricity pricing from Eskom had on Wits University’s electricity bill. This code has
been verified and matches that of online bills (MOL, 2008). There is however one
discrepancy, the peak complex demand (see below) is being charged on the peak for
the billing month as opposed to that set out in the tariff guidelines provided by City
Power (City of Johannesburg, 2008). A summary of these guidelines is provided in
Appendix E. This code has further been adjusted to incorporate the effects the CSP
alternatives have on the cost of electricity.
Wits University is billed yearly for the period 1 July to 30 June. Because the demand
charge is calculated using twelve months worth of data (as outlined in Appendix E),
to analyse the bill, data from July 2006 - June 2007 as well as July 2007 – June 2008
are used. This can be done for each of Wits University’s six campuses.
96
The three output files from the design analysis of the twelve different CSP systems
are inputted into the code (the solar output, the storage output and the hybrid
output for each of the twelve configurations). These files merely contain the actual
kW of electricity production on a half hourly basis. This data was then integrated
with the electricity usage at Wits University and its effect on the electricity
consumption and bill was calculated.
Wits University’s half hourly consumption data has been arranged into an Excel
spreadsheet that contains the real, reactive and complex usage demands for each
year. The real and complex power demand is then adjusted in Matlab to take into
account the effect of the solar-electricity generation.
Complex power
To calculate the rates for the complex power, 80 % of the average of the three
largest peaks of the preceding twelve months for each month is calculated. This
value is then compared to find the greater of the measured demand for the month of
interest and the demand of 70 kVA (see Appendix E). This process is also performed
for the peak demand after the effect of the solar power integration.
The average power factor for Wits University West campus over two years was
found to be 0.9. This power factor that is measured when feeding power into the
University is assumed to be the same power factor the CSP systems will operate
with when feeding electricity into the university’s grid.
Reactive power
To determine the billable reactive energy, a charge is made on the kVARh supplied
in excess of 30% (0,96PF) of kWh for each month. The reactive energy costs are
calculated on a monthly and yearly basis by summing the applicable billable reactive
energy.
97
Real power
The real usage is summed for the month to find the cost of real electric power for
each month. The different summer/winter charge rates are applied where
applicable. Again, this is also done for the usage that incorporates the effect of the
addition of solar power/hybrid generation. These are then summed to find the
yearly billing data.
Total Bill
A surcharge of 2% (see Appendix E) of the total energy costs (sum of the electricity
usage cost, peak demand cost, billable reactive energy cost and the constant service
charge) is added to the cost. Finally the total energy bill is determined by adding a
VAT (14%) charge to the total which includes the surcharge.
LEC
The real LEC for each of the CSP systems is also inputted into the code. The actual
cost of the CSP electric generation per kWh is found by totalling the actual cost per
kWh of solar electricity produced, using the LEC, less the savings made on the
complex power and actual power consumption. From this, a monthly value for a
‘Wits’ LEC can be found and averaged for the year.
Capacity Factor
Each of the systems has been designed for an average capacity factor for the year.
Depending on seasonal variations, the actual capacity factor for each month will
vary from this average value. The code finds the average capacity factor for each
month to show where production is maximised.
98
7 RESULTS
7.1 Initial Design Results
Table 7-1 shows the results for the required energy flows for the reference plants on
yearly basis. These energy flows, at the set capacity factors, were used to calculate
the resulting plant aperture areas shown in Table 7-2 and the aperture and total
areas are depicted in Figure 7.1. The plant areas for the 480 kW(e) systems are
shown in Figure 7.2.
Table 7-1: Thermal Energy Flow
Design Electric
# Collector Type Power Cycle Solar CF Enet [Wh/a] Qtherm [Wh/a] Qsf [Wh/a]
Output [kW(e)]
1&3 Parabolic Trough Steam Cycle 120 20% 210,240,000 592,225,352 766,428,395
2 Parabolic Trough Steam Cycle 120 30% 315,360,000 592,225,352 1,129,332,240
4&6 Parabolic Trough ORC 120 20% 210,240,000 914,086,957 1,182,965,566
5 Parabolic Trough ORC 120 30% 315,360,000 914,086,957 1,743,099,761
7&9 CLFR Steam Cycle 120 20% 210,240,000 592,225,352 778,318,095
8 CLFR Steam Cycle 120 30% 315,360,000 592,225,352 1,148,019,191
10&12 CLFR ORC 120 20% 210,240,000 914,086,957 1,201,317,060
11 CLFR ORC 120 30% 315,360,000 914,086,957 1,771,942,664
Table 7-2: Aperture Areas
# Collector Type Power Cycle Aa [m2]

1&3 Parabolic Trough Steam Cycle 841
2 Parabolic Trough Steam Cycle 1,239
4&6 Parabolic Trough ORC 1,297
5 Parabolic Trough ORC 1,912
7&9 CLFR Steam Cycle 1,090
8 CLFR Steam Cycle 1,607
10&12 CLFR ORC 2,480
11 CLFR ORC 1,682
99
9,000
Solar Field Area
Total Plant Area
8,000
7,000
6,000
Area [m 2]
5,000
4,000
3,000
2,000
1,000
-
PT PT, Stor PT, PT, ORC PT, Stor, PT, CLFR CLFR, CLFR, CLFR, CLFR, CLFR,
Hybrid ORC Hybrid, Stor Hybrid ORC Stor, Hybrid,
ORC ORC ORC
Technology Option
Figure 7.1: Reference Plant Areas (120 kW(e))
20,000.00 Solar Field Area

Total Plant Area
15,000.00
Area [m 2]
10,000.00
5,000.00
-
CLFR, ORC CLFR, Stor, ORC CLFR, Hybrid, ORC
Technology Option
Figure 7.2: Plant Area for CLFR, ORC technologies (480 kW(e))
7.2 Initial Financial Results

Table 7-3 below shows the financial results that are based on the average economic
data in Table 5-13. It is this data that has been used to determine the real LEC for
the solar electricity production. Figure 7.3 below the table is a graph that plots the
twelve alternatives against each other. This figure also includes the specific
investment requirements of the alternatives.
100
Table 7-3: Economic Results
PT, Hybrid, CLFR, CLFR, Stor, CLFR,

System PT PT, Stor PT, Hybrid PT, ORC PT, Stor, ORC CLFR CLFR, Stor CLFR, ORC
ORC Hybrid ORC Hybrid, ORC
Size of field [m2] 841 1,239 841 1,297 1,912 1,297 1,090 1,607 1,090 1,682 2,481 1,682
Total Area 3,362 4,954 3,362 5,189 7,647 5,189 1,676 2,473 1,676 2,587 3,817 2,587
Investment Solar
R 3,495,459 R 5,150,559 R 3,495,459 R 5,395,165 R 7,949,776 R 5,395,165 R 2,189,055 R 3,228,857 R 2,189,055 R 3,378,759 R 4,983,670 R 3,378,759
Field
Investment Power
R 1,901,368 R 1,901,368 R 1,901,368 R 1,744,356 R 1,744,356 R 1,744,356 R 1,901,368 R 1,901,368 R 1,901,368 R 1,744,356 R 1,744,356 R 1,744,356
Block, BOP
Investment Storage R0 R 656,199 R0 R0 R 1,012,829 R0 R0 R 656,199 R0 R0 R 1,012,829 R0
Investment Land R0 R0 R0 R0 R0 R0 R0 R0 R0 R0 R0 R0
Contingencies R 1,619,048 R 2,312,438 R 1,619,048 R 2,141,856 R 3,212,088 R 2,141,856 R 1,227,127 R 1,735,927 R 1,227,127 R 1,536,934 R 2,322,256 R 1,536,934
Sum Total Equipment
R 5,396,827 R 7,708,126 R 5,396,827 R 7,139,521 R 10,706,961 R 7,139,521 R 4,090,423 R 5,786,423 R 4,090,423 R 5,123,115 R 7,740,855 R 5,123,115
Costs
Total Including
R 7,015,875 R 10,020,564 R 7,015,875 R 9,281,377 R 13,919,049 R 9,281,377 R 5,317,550 R 7,522,350 R 5,317,550 R 6,660,049 R 10,063,111 R 6,660,049
indirect Costs
Actual O&M Costs R 201,706 R 288,091 R 201,706 R 185,628 R 278,381 R 185,628 R 152,880 R 216,268 R 152,880 R 133,201 R 201,262 R 133,201
Annual Financing
R 814,397 R 1,163,179 R 814,397 R 1,077,375 R 1,615,713 R 1,077,375 R 617,257 R 873,189 R 617,257 R 773,094 R 1,168,118 R 773,094
and insurance Costs
Annual Fuel Costs R 587,111 R 904,612 R 587,111 R 904,612
Actual Net Elec [kWh] 210,240 315,360 568,699 210,240 315,360 568,699 210,240 315,360 568,699 210,240 315,360 568,699
Solar Net
210,240 315,360 210,240 210,240 315,360 210,240 210,240 315,360 210,240 210,240 315,360 210,240
Electricity(Joburg)
Fossil net Electricity
- - 358,459 - - 358,459 - - 358,459 - - 358,459
[kWh]
Fossil thermal energy
- - 1,009,744 - - 1,558,518 - - 1,009,744 - - 1,558,518
required [kWhth]
- - 3,635 - - 5,611 - - 3,635 - - 5,611
[GJ]
- - 4,544 - - 7,013 - - 4,544 - - 7,013
input [GJ]
Total Costs [R] R 7,015,875 R 10,020,564 R 7,015,875 R 9,281,377 R 13,919,049 R 9,281,377 R 5,317,550 R 7,522,350 R 5,317,550 R 6,660,049 R 10,063,111 R 6,660,049
Specific Investment
R 58,466 R 83,505 R 58,466 R 77,345 R 115,992 R 77,345 R 44,313 R 62,686 R 44,313 R 55,500 R 83,859 R 55,500
(R/kW)
LEC -Real (R/kWh) R 4.83 R 4.60 R 2.82 R 6.01 R 6.01 R 3.81 R 3.66 R 3.45 R 2.39 R 4.31 R 4.34 R 3.18
101
R 7.00
R 6.00 PT, ORC PT, Stor, ORC
R 5.00
PT
PT, Stor
CLFR, ORC CLFR, Stor, ORC
R 4.00
PT, Hybrid, ORC
CLFR
CLFR, Stor PT
CLFR, Hybrid, ORC PT, Stor
R 3.00 PT, Hybrid
PT, Hybrid
Real LEC [R/kWhe]

PT, ORC
CLFR, Hybrid PT, Stor, ORC
PT, Hybrid, ORC
R 2.00
CLFR
CLFR, Stor
CLFR, Hybrid
CLFR, ORC
R 1.00
CLFR, Stor, ORC
CLFR, Hybrid, ORC
R 0.00
R 0.00 R 20,000.00 R 40,000.00 R 60,000.00 R 80,000.00 R 100,000.00 R 120,000.00 R 140,000.00
Specific Investment [R/kW]
Figure 7.3: Economic Results (120 kW(e) reference systems)
102
7.3 Matlab Modelling
Chosen Technologies for further investigation

Energy modelling of the twelve reference systems was carried out and a second
perspective model was put together that ranks the technologies against one another
using the same factors discussed in Section 4. This can be found in Appendix I. This
was used as an indicator and is not followed directly. The three options chosen for
further investigation are given below with reasons that follow. Modelling of these
systems was done at a 480 kW(e) capacity.
• CLFR with ORC

• CLFR with ORC making use of storage
• CLFR with ORC with hybridisation using natural gas.
Solar Field
In terms of urban distributed power generation the solar field that shows the most
potential is the CLFR configuration. It is the most compact and offers the smallest
plant area for a set electric output. The CLFR option also offers large infrastructure
savings. The solar field also requires less water for cleaning and considering the
sustainability of the water supply in a rapidly growing city such as Johannesburg,
this benefit is favoured.
Power Cycle
In terms of the potential power cycle, an ORC is recommended for further
investigation. ORC plants are noted to have less demanding operating assistance
because they are capable of automatic start-up, safe shutdown, and regulation with
varying solar conditions. It is more common to use ORC systems for small scale
generation and because the ORC systems can operate at lower temperatures; the
efficiency of the solar field is less important. This allows room for savings in solar
field costs.
103
Energy modelling
The following discusses the modelling that was used for the 120kW(e) reference
systems as well as the chosen design alternatives at 480 kW(e). For simplicity, only
the results for the three alternatives discussed above have been displayed.
Energy Flow
The hourly energy flow from the initial DNI collected by the solar field to the electric
generation has been tracked and represented in Figure 7.4, Figure 7.6 and Figure 7.8
(Note different scales used). The graphs respectively represent the hourly energy
flow for a typical year for the three alternatives discussed above. Figure 7.5, Figure
7.7 and Figure 7.9 show the average hourly energy for a 24 hour period for each of
the systems.
Effects on the Usage

Figure 7.10, Figure 7.12 and Figure 7.14 each represent the effect that the three
alternatives will have on the electricity usage at the University (for a typical billing
year – June to July). The black section on the graphs represent the solar generation.
It can be seen that the largest effect on the usage is with the use of hybridisation
because it has the highest capacity factor (Figure 7.14).
Figure 7.11, Figure 7.13 and Figure 7.15 show the effects of the three alternatives on
the daily consumption. The day chosen is a typical day in winter during term time
(17 June 2008). With this scale used the effect of the three alternatives can be seen
in greater detail. Figure 7.13 shows the effects of the storage system coming on line
at 18h00. The electricity generated using hybridisation can be clearly seen as a
constant input between the hours 07h00 and 20h00 in Figure 7.15.
104
Figure 7.4: Hourly Energy Flow for CLFR, ORC (480 kW(e)) Figure 7.6: Hourly Energy Flow for CLFR, ORC, with Figure 7.8: Hourly Energy Flow for CLFR, ORC, with
Storage (480 kW(e)) Hybridisation (480 kW(e))
Figure 7.5: Average Energy Flow for CLFR, ORC (480 Figure 7.7: Average Energy Flow for CLFR, ORC,
with Storage (480 kW(e)) Figure 7.9: Average Energy Flow for CLFR, ORC, with
kW(e)) Hybridisation (480 kW(e))
105
Figure 7.10: West Campus Power Usage - CLFR, ORC (480 Figure 7.12: West Campus Power Usage - CLFR, Figure 7.14: West Campus Power Usage - CLFR, ORC,
kW(e)) ORC, with Storage (480 kW(e)) with Hybridisation (480 kW(e))
Figure 7.11: West Campus Power Usage - CLFR, ORC (480 Figure 7.13: West Campus Power Usage - CLFR, Figure 7.15: West Campus Power Usage - CLFR, ORC,
kW(e)) ORC, with Storage (480 kW(e)) with Hybridisation (480 kW(e))
106
Energy Usage and Production
Table 7-4, Table 7-5 and Table 7-6 show the monthly energy usage as well as the total
energy produced using each of the three alternatives.
Billing Results
Table 7-7, Table 7-8 and Table 7-9 are the billing results for each of the three systems
under analysis. A summary of this data is given in Table 7-10. A summary of the billing
results for the reference plants is given in Appendix J. As mentioned in Section 5.6.3, the
inclusion of the REFIT for a CSP application has been included merely as an indicator for
what the nominal LEC would be if the electricity were to be sold.
LEC and Payback

The LEC and payback for the three alternatives analysed are given in Figure 7.16 and
Figure 7.17. Figure J1 and Figure J2 in Appendix J show the same results for 120 kW(e)
reference plants.
107
Table 7-4: Energy Results using CLFR, ORC
CLFR, ORC
Year 2007 2007 2007 2007 2007 2007 2008 2008 2008 2008 2008 2008
Month 7 8 9 10 11 12 1 2 3 4 5 6
Energy Consumption [kWh] 1,441,049 1,341,473 1,097,243 1,143,128 1,017,735 668,810 737,151 952,793 1,022,896 986,482 1,249,898 1,263,414
Peak Complex Power [kVA] 3,043 2,767 2,453 2,372 2,340 1,800 1,870 2,421 2,415 2,321 2,822 2,837
Billable Complex Power [kVA] 3,043 2,767 2,552 2,552 2,552 2,552 2,552 2,552 2,552 2,552 2,822 2,837
Reactive Energy [kVAR] 457,161 455,943 510,416 488,742 501,856 377,258 413,431 527,204 530,937 494,766 495,731 406,323
Billable Reactive Energy [kVAR] 24,846 53,501 181,243 145,804 196,535 176,615 192,285 241,366 224,068 198,821 120,761 27,299
Solar Energy Generated [kWh] 97,475 93,336 79,003 62,359 50,423 57,750 49,103 51,303 65,737 68,743 95,463 88,162
Capacity Factor 0.28 0.27 0.23 0.18 0.15 0.17 0.14 0.15 0.19 0.20 0.28 0.26
Table 7-5: Energy Results using CLFR, ORC with Storage
CLFR, ORC
Year 2007 2007 2007 2007 2007 2007 2008 2008 2008 2008 2008 2008
Month 7 8 9 10 11 12 1 2 3 4 5 6
Capacity Factor 0.43 0.41 0.35 0.29 0.23 0.27 0.25 0.24 0.29 0.30 0.42 0.39
108
Table 7-6: Energy Results using CLFR, ORC with Hybridisation
CLFR, ORC
Year 2007 2007 2007 2007 2007 2007 2008 2008 2008 2008 2008 2008
Month 7 8 9 10 11 12 1 2 3 4 5 6
Capacity Factor 0.56 0.56 0.54 0.56 0.54 0.56 0.56 0.52 0.56 0.54 0.56 0.52
Table 7-7: Billing Results using Normal CLFR, ORC
CLFR, ORC
Year
2007 2007 2007 2007 2007 2007 2008 2008 2008 2008 2008 2008
Month
7 8 9 10 11 12 1 2 3 4 5 6
Consumption Costs Normal [R]
499,756 465,223 257,304 268,064 238,659 156,836 172,862 223,430 239,869 231,330 433,465 438,152
Consumption Costs with Solar Generation [R]
465,951 432,854 238,777 253,440 226,835 143,294 161,347 211,399 224,454 215,210 400,358 407,578
Demand Costs Normal [R]
245,489 223,188 199,630 199,630 199,630 199,630 199,630 199,630 199,630 199,630 227,646 228,822
Demand Costs with Solar Generation [R]
229,796 223,188 198,109 198,109 198,109 198,109 198,109 198,109 198,109 198,109 204,708 225,746
Billable Reactive Energy Costs [R]
1,523 3,280 11,110 8,938 12,048 10,826 11,787 14,796 13,735 12,188 7,403 1,673
Total Costs Normal [R]
869,730 805,686 545,630 555,616 525,040 428,477 448,229 510,528 528,410 516,681 778,736 778,891
Total Costs with Solar Generation [R]
812,175 768,048 522,319 536,843 509,522 410,960 433,070 494,769 508,716 496,167 713,567 739,763
Total Costs including Costs of Solar Generation [R]
1,232,291 1,170,325 862,821 805,611 726,845 659,862 644,704 715,885 792,043 792,451 1,125,012 1,119,740
Wits Solar LEC [R/kWh]
3.80 3.96 4.06 4.05 4.05 4.05 4.04 4.05 4.05 4.05 3.72 3.93
109
Table 7-8: Billing Results CLFR, ORC with Storage
CLFR, ORC
Year 2007 2007 2007 2007 2007 2007 2008 2008 2008 2008 2008 2008
Month 7 8 9 10 11 12 1 2 3 4 5 6
Consumption Costs Normal [R] 499,756 465,223 257,304 268,064 238,659 156,836 172,862 223,430 239,869 231,330 433,465 438,152
Consumption Costs with Solar Generation [R] 448,311 415,930 228,534 244,902 219,689 134,996 152,961 203,629 216,214 206,854 383,455 390,857
Demand Costs Normal [R] 245,489 223,188 199,630 199,630 199,630 199,630 199,630 199,630 199,630 199,630 227,646 228,822
Demand Costs with Solar Generation [R] 217,698 215,417 195,613 195,613 195,613 195,613 195,613 195,613 195,613 195,613 201,688 225,746
Billable Reactive Energy Costs [R] 1,523 3,280 11,110 8,938 12,048 10,826 11,787 14,796 13,735 12,188 7,403 1,673
Total Costs Normal [R] 869,730 805,686 545,630 555,616 525,040 428,477 448,229 510,528 528,410 516,681 778,736 778,891
Total Costs with Solar Generation [R] 777,596 739,332 507,505 524,012 498,310 398,410 420,416 482,831 496,232 483,548 690,401 720,320
Total Costs including Costs of Solar Generation [R] 1,421,394 1,356,205 1,039,955 952,669 849,397 802,604 788,734 849,303 934,026 936,543 1,316,237 1,312,195
Wits Solar LEC [R/kWh] 3.81 3.94 4.07 4.06 4.06 4.06 4.06 4.06 4.07 4.07 3.81 3.97
Table 7-9: Billing Results using CLFR, ORC with Hybridisation
CLFR, ORC
Year 2007 2007 2007 2007 2007 2007 2008 2008 2008 2008 2008 2008
Month 7 8 9 10 11 12 1 2 3 4 5 6
Consumption Costs Normal [R] 499,756 465,223 257,304 268,064 238,659 156,836 172,862 223,430 239,869 231,330 433,465 438,152
Consumption Costs with Solar Generation [R] 432,671 398,138 213,405 222,702 194,760 111,474 127,500 180,995 194,507 187,432 366,380 375,395
Demand Costs Normal [R] 245,489 223,188 199,630 199,630 199,630 199,630 199,630 199,630 199,630 199,630 227,646 228,822
Demand Costs with Solar Generation [R] 202,846 215,417 175,540 175,540 175,540 175,540 175,540 175,540 175,540 175,540 195,290 193,158
Billable Reactive Energy Costs [R] 1,523 3,280 11,110 8,938 12,048 10,826 11,787 14,796 13,735 12,188 7,403 1,673
Total Costs Normal [R] 869,730 805,686 545,630 555,616 525,040 428,477 448,229 510,528 528,410 516,681 778,736 778,891
Total Costs with Solar Generation [R] 742,139 718,644 466,573 474,857 445,983 347,718 367,470 433,172 447,652 437,624 663,105 664,448
Total Costs including Costs of Solar Generation [R] 1,357,278 1,333,783 1,061,869 1,089,997 1,041,279 962,857 982,609 1,008,625 1,062,791 1,032,920 1,278,245 1,239,900
Wits Solar LEC [R/kWh] 2.61 2.79 2.82 2.82 2.82 2.82 2.82 2.81 2.82 2.82 2.67 2.64
110
Table 7-10: Summary for CLFR, ORC Technologies at 480 kW(e)

Total Electricity Consumption [kWh] 12,922,073 12,922,073 12,922,073
Total Solar Electricity Generated [kWh] 858,856 1,341,094 2,277,600
Yearly Bill [R] 7,291,654 7,291,654 7,291,654
Total Bill [R] (incl. cost of Solar) 10,647,589 12,559,262 13,452,154
Extra cost for Solar [R/year] 3,355,935 5,267,608 6,160,499
Cost Saved on Bill [R/year] 345,735 552,740 1,082,269
Real LEC [R/kWh] 4.31 4.34 3.18
Wits LEC [R/kWh] 3.98 4.00 2.77
Average Capacity Factor 0.21 0.32 0.55
Total Investment [R] 26,640,198 40,252,445 26,640,198
Payback [years] 77 73 25
- - -
Nominal LEC [R/kWh] (with REFIT) 1.93 1.95 2.01
Extra cost for Solar [R/year] (with REFIT) 1,595,280 2,518,365 4,431,210
Payback [years] (with REFIT) 12.65 12.19 9.48
LEC Solar LEC

5.0 Real LEC
4.31 4.34 Nominal LEC (with FIT)

4.5
3.98 4.00
4.0
3.5
3.18
LEC [R/kWh(e)]
3.0 2.77
2.5
1.93 1.95 2.01
2.0
1.5
1.0
0.5
-
Technology
Figure 7.16: LEC Results for CLFR, ORC technologies (480 kW(e))
111
Payback Payback
90
Payback - with FIT
80 77.1
72.8
70
60
Payback [Years]
50
40
30 24.6
20
12.6 12.2
9.5
10
-
Technology option
Figure 7.17: Payback Results for CLFR, ORC Technologies (480 kW(e))
112
8 DISCUSSION
8.1 Suitability of Design Approach

The scope of work only included the analysis of solar-thermal technologies. This
excluded conventional diesel generators and other renewable technologies such as
photovoltaic systems and biogas digesters. A comparison of these technologies to
the solar systems investigated would be useful in showing the cost of distributed
power options as opposed to grid-connected power. It is therefore recommended
that these be included in future investigations.
Consistency in the data analysis was very important, specifically in terms of the
costing of the systems. Because they are all on different levels of maturity, absolute
cost data are difficult to estimate, however the relative distribution of the different
cost items is considered to be well estimated by the approach followed. Because of
the inherent uncertainty and variability of costs, the modelled technical
performance of the system designs was more accurate than the costing analysis,
however through the methodology followed, certain discrepancies were identified.
These are further elaborated.
Scaling effects
The technical and specific cost data from the studies done on utility scale plants
(>50 MW(e)) have been included in the analysis. The scaling of the technologies is
not linear, as suggested by the scaling methods used and this will lead to several
issues, most importantly, where the data provided is based on the economies of
scale of the larger plants. Whether these methods are appropriate for scaling down
below 1 MW(e) is questionable and should be investigated further. The technical
aspects of some of the systems, such as the efficiency of the power block, may not
scale as suggested. Freepower (2008) claim efficiencies of up to 22% at 270 dC. The
scaling of the steam cycle generators at such small capacities usually results in a
greater drop in efficiency. These lower efficiencies will therefore require greater
solar field areas in order to deliver the required thermal energy.
113
Whether these methods are also appropriate for cost scaling, as with the technical
performance, is also questionable. The scaling effects of the costs will also be
manufacturer and country specific and in order to calculate absolute cost data,
direct quotes from the various manufacturers as well as a full life-cycle cost analysis
will need to be performed.
Costing
In finding the present value of the costs from the multiple sources, the Chemical
Engineering Plant Cost index was used. The index is adjusted according to the dollar
(USD$) increase in the price of goods. The index also takes into account labour
effects which are inherently different at all locations. However, the general cost
inflation of the materials and services is what is needed, and the results of this are
used in the comparative analysis. Because of the variability of financial parameters,
such as interest rates, inflation, incentives and tariffs, and exchange rates, which
change on a daily basis, the importance of relative costs is again emphasised.
Levelised electricity cost (LEC)

The LEC approach was chosen for the financial analysis because of its comparative
advantages. Other possible decision making tools include the Net Present Value and
Internal Rate of Return approach. These criteria are used mainly for specific
investment decisions and depend on specific financing policies which can differ
dramatically between institutions. An example where the LEC approach is most
valuable is when comparing technologies such as the plants with storage, which
produce more electricity than plants under solar-only operation, which will have
greater installation costs (due to larger solar-field size). Whether the benefits of this
added generation will outweigh the added capital costs can be found by levelising
the data which will then make a level platform for comparison. (Storage is however
a design decision and is, most often than not, utilised for reasons other than
economic, such as the timing of dispatch).
114
The factors that affect the financial feasibility of the installation of a distributed CSP
generation system have been outlined through the cost analysis. Several factors
affecting the cost of electricity have been identified and are outlined below.
Certain discrepancies in using the cost data from various studies were also
identified. For example, the cost of technology will play a significant role in
determining the LEC. Whether the technologies are manufactured locally or
imported, like the receivers and turbines, which are specialised items, costs will
differ from those used in this report. At the same time costs derived from other
aspects of the installation and running of the plant will be less than estimated. An
example of this is the O&M costs, which are expected to be lower in developing
countries. Becker et al. (2000) suggest that O&M costs in developing countries will
be approximately 15% less than those found in developed nations such as Europe.
Other costs, such as shipping costs to South Africa and international professional
fees will increase costs and, the for the purposes of this study, these differences
mentioned, have been assumed to approximately balance.
Different institutions have different capital structures and obligations and the value
of such an investment will then determine the appropriate financing measures.
Whether the project is financed entirely by debt or through the use of equity will
determine the fixed charge rate (fcr) and hence financing costs. This will have a
significant effect on the LEC. Unlike in a lot of the European countries, there are few
opportunities in South Africa to invest in CSP systems (especially of this size).
Financing the project through the use of debt may be difficult because there little
financial return on the capital investment. If the project is then financed through the
use of equity, financing costs will therefore translate into an opportunity cost, or the
loss in income from an investment of equivalent capital proportion. This is the
reasoning behind the choice of a money market interest rate for the fixed charge
rate, in Section 5.6.
115
The installation of a renewable energy system will provide numerous intangible
benefits that cannot be measured simply through indicators such as the LEC. These
benefits will also reach far beyond those described in the introduction (economic,
energy security, climate change) and will often be immeasurable. An example of
such benefits may include customer satisfaction, with the view of ethical
management, when a firm chooses to ‘go green’. Depending on the case, a lot of
pressure is placed on certain commercial institutions to consider their
environmental policies. This can also have a negative side - customers may view this
as negligence on management’s behalf because funds are ‘unnecessarily spent’ and
the future growth or prosperity of the firm is questionable. Energy security in
commercial and industrial applications often transpires into economic issues. With
the power cuts experienced in 2008, the financial and productive losses experienced
were, in many cases, far greater than the cost of electricity produced from the
various CSP systems investigated here. It is therefore up to each institution to put a
price on these benefits.
Plant design and performance

The method used to calculate the required plant areas for the design configurations
was based on common annual design capacity factors. This method, as opposed to
designing a plant according to a design point peak DNI, was chosen because, for the
evaluation of the technologies with respect to their effect on Wits University, the
same electric production between plants was necessary for level comparison. It is
recommended that the three chosen configurations be further designed according to
the design point peak DNI. This design procedure, especially the optical part of the
system (concentrator field arrangement and size, receiver aperture, orientation and
height) should also be cost-optimised.
For the energy modelling a Matlab code was written as a shell that is able to analyse
hourly DNI data from any location. The code has been programmed to analyse the
thermal energy conversion of the twelve design alternatives listed in Section 5.4.
The model is easily adaptable to any CSP application. The analysis of hourly DNI
116
data was necessary because Wits University’s electricity bill is determined from
half-hourly usage. The effects that the electricity generated from the CSP
applications have on the total bill are then calculated with reasonable accuracy.
In order to understand the plant performance, the methodology behind the various
studies was verified. Because this report considers the feasibility of CSP systems in
terms of basic cost and performance data, a full thermodynamic design/analysis was
not required. This level of analysis is satisfactory in order to conclude whether
further research needs to be undertaken and at what level.
8.2 Results
In terms of the technical viability, the average solar resource for a typical year in
Johannesburg is sufficient for power production but when analysed on an hourly
basis, it is seen that the resource is very intermittent. This is due to the amount of
cloud cover experienced, especially in the summer months (Figure 7.4). Solar-only
operation is therefore unsuccessful in generating cost savings benefits by cutting
down the peak usage at Wits.
Figure 7.1 shows the plant areas as well as the solar field areas of the various design
options. The CLFR technologies have larger solar field (aperture) areas due to the
lower efficiency of the field but because they are more compact, areas can be
smaller than those for parabolic troughs. Because of the complete installation into a
space constrained environment it is the total plant area that is deemed the most
important here. Comparing the various plant sizes to the space available at Wits
University, there are very obvious options for integration. Through modulation,
several plants can be installed at various locations. The option of introducing a plant
above one of the parking lots will also create shading benefits for cars, and if
installed on a building roof, similar shading benefits are found.
117
According to Figure 7.3, which shows the real LEC (LEC without the savings from
generation at Wits University and excluding the REFIT), the hybrid options result in
a significantly lower LEC. The CLFR which uses a normal steam cycle with hybrid
generation came out financially the most feasible option. The CLFR collector that
uses an ORC with hybrid mode resulted in the third cheapest option. This shows the
effect that the specific investment for the ORC technology has on the LEC. Here, in
order to provide enough thermal energy to the ORC (which has lower conversion
efficiency than a normal steam cycle), a larger solar field is required. It is the higher
costs attributed to this field as well as the higher costs demanded from the power
block that make the ORC alternatives more expensive than their steam cycle
counterpart. Even the lower operating costs incurred by the ORC do not compensate
for the higher investment costs.
Storage
It is interesting to note that, in Figure J1 (Appendix J), the storage options actually
bring the LEC down in most cases whereas for the chosen CLFR, ORC system, the
price of electricity increases. The combination of CLFR with the ORC electric
generation causes the efficiency to be significantly low enough to cause the cost of
storage to be higher. This higher required thermal input results in infrastructure
costs that cause the price of electricity to be higher than the option that doesn’t
incorporate storage. A cost optimisation for the amount of heat storage and solar
multiple is recommended to find the optimal storage level.
Out of the three 480 kW(e) options investigated, the hybrid option was the only
system that successfully brought down the peak electricity demand (Figure 7.15).
The option with no storage wasn’t able to eliminate the evening peak (Figure 7.11),
and the option with storage (Figure 7.13) eliminated the evening peak but the
morning peak remained, which is almost equivalent to the evening peak. To
successfully eliminate the morning peak as well, the storage capacity will have to be
greatly increased and dispatched the following morning. The use of storage is
therefore effective in bringing down the evening peak load when sunlight is
118
available on the same the day. In terms of financial feasibility and space utilisation,
energy storage is unlikely to be feasible. Storage for the systems was designed to
provide 3 hours of electricity to match the evening peak load, which is appropriate
for this investigation but it is recommended that the storage be sized optimally in
terms of costs as well as for reliable dispatch. Benefits, which will be gained by the
evening dispatch through storage, include energy security and the possibility of a
higher cost saving with the possible introduction of a Time of Use (TOU) charge.
Hybridisation
Including hybridisation into the design configurations resulted in the lowest cost of
electricity. These results however do not take into account the added infrastructure
costs needed, such as linking of a gas line and boiler costs. The price is therefore
underestimated but still assumed to be the cheapest option. It is important to note
here that the hybrid options, that yield the lowest LEC, include the generation of
electricity from natural gas and are not representative of a solar-only LEC. This
value may be misleading if not considered in context. The hybrid options are
designed to run off a capacity factor of 54%. If the solar-only systems incorporated
the cost of electricity from City Power and a new LEC was determined where the
capacity factor of these systems was increased to 54% (from 20 and 30%), the
solar-only systems would in fact be cheaper. However, the analysis here was done to
determine the cost of electricity that can be generated, in this case, at Wits
University. It is energy security that is the priority and the cost of this security that
is determined.
Arising from this point is the possibility of creating a natural gas-only generation
system which would offer even further reductions in the cost of electricity
generated. This would eliminate the reliance on the intermittent solar resource as
well as provide better energy security, similar to diesel generators, which have
become very popular in industry. The advantage of electricity production from
piped natural gas at Wits University is that the fuel does not need to be transported
like diesel for example. So the implementation of such distributed CSP systems will
119
perhaps find better feasibility in off-grid communities where the cost of extending
the grid and transporting liquid fuels is expensive. The renewable energy argument
has however been debated for decades and the aim of this report is to investigate
the feasibility of using the solar resource for electricity generation through CSP.
Also a technical issue with the hybrid options is the emissions from the combustor.
Wits University and urban areas in general are relatively dense. The pollution from
cars in most cities is a problem and to add a co-generation plant in an urban
environment will only make the situation worse. If the system is located on a roof-
top this will be less of a problem to immediate bystanders. One of the reasons for
installing such a renewable energy system is to combat emissions and mitigate
climate change, so the hybrid options from this point of view are less viable.
8.3 Recommendations for Implementation
Financing options – REFITs, CDM and TRECs

The main motivation of energy security for the CSP application investigated here, is
again restated. The sale of electricity under the REFIT, especially for large scale
commercial applications, can make CSP feasible in South Africa. However an
application that satisfies a certain energy load, through off-grid, distributed power,
will not be able to sell electricity to the Renewable Energy Purchasing Agency
(REPA) itself (under current conditions (NERSA, 2008)).
A cap on the amount of renewable energy REPA is willing to purchase is expected.

However, after the introduction of the REFIT, it is also expected that there will be an
increased market demand for clean energy. This will make it more likely that certain
institutions will be interested in signing off-grid PPAs whereby clean electricity is
supplied to them directly. This is expected because of international pressure for
climate change mitigation as well as energy security. The CSP application
investigated would generate energy to be used by Wits University itself and
therefore will not sell electricity to external parties. The effect of a REFIT was
120
however analysed in order to indicate results for a nominal LEC should an
institution (in this case Wits University) be willing to pay the market related price
(CSP REFIT) for CSP generated electricity.
At the time of writing, a CSP REFIT of R2.10/kWh (Fin24, 2009) was announced by
NERSA. The value of this tariff is only slightly higher than the one used in the
analysis (R2.05/kWh) and should greatly boost investment in the sector, to reach
the renewable energy targets set out in the RE White Paper (DME, 2003).
The fact that Eskom’s price of electricity in essence, does not allow for the recovery
of all the prudently incurred costs and the building of reserves to sustain the current
asset base; nor does it support the capital expansion, especially with their intention
of doubling capacity by 2026 (Eskom (b), 2009). This indicates heavy tariff
increases in the future. Price parity will be reached in the near future where the cost
of CSP technology will be more competitive than traditional fossil-fuelled power
generation. It is recommended that financial predictions be performed with respect
to the South African electricity tariffs (as well as other distributed sources of power)
to aid investors in decision-making.
Other financing options mentioned in Section 5.6.3 include financing through the
Clean Development Mechanism (CDM) and Tradable Renewable Energy Certificates
(TRECs). Certain disadvantages with these sources of finance have been discussed,
and with the instability of the world’s economy, the reliance on such mechanisms
makes firm investment decisions debateable (IETA, 2009). These financing options
have not been incorporated into the analysis and it is recommended that these be
further investigated. These should be incorporated into a full discounted cash flow
to predict the future costs of electricity. Other factors to be considered is carbon
taxing that, if implemented in South Africa, would influence the price of electricity.
121
Technical Recommendations
scale CSP systems are currently higher than conventional fossil fuel options.
Exploiting the full potential of high efficiencies and economies of scale of plants with
power levels above 50 MW(e), a very high investment cost is required. Although,
Eskom is currently pursuing the installation of a 100 MW(e) plant, independent
power producers may find such scales intimidating. With the introduction of these
technologies at lower power levels, cost savings with the incorporation of other
design options should be pursued. These options are now discussed.
Culwick (2008) makes an interesting point about the current energy use in South
Africa. Currently 30% of South Africa’s domestic energy usage is to heat water (80%
of which is electrical energy). Electricity generation is a high quality use of solar
power, necessitating the requirement for concentrating collectors and to use
electricity to heat water at low temperatures is therefore a waste of the high quality
source. According to Aitken (2003) one square metre of surface area can deliver 100
W of peak electrical power with PV technology. Comparing this to CSP generation,
one square metre of mirror can also deliver about 100 W of peak electrical energy.
However, one square metre of intercepted solar energy can also deliver 300 W of
thermal power for heating domestic water, displacing 300 W of electric water
heating.
Therefore, by considering the matching of the energy source to the energy use,
certain design options result. These may include the incorporation of waste heat
from the power cycle into Wits University’s hot water or heating and air-
conditioning systems. This will have a significant impact on the cost of electricity,
and is definitely worth pursuing. It will be necessary to investigate how much of the
electric energy is used for hot water and heating. Perhaps the best implementation
approach, which would also kick-start CSP research, is by first installing a solar field
that is used solely to collect heat for various systems such as hot-water or HVAC
requirements. An advantage of installing such fields on top of buildings is the
122
shading they will provide, which will bring down the air-conditioning demand.
Further, 5% of electricity consumed in South Africa is used for process heating in
the commercial and industrial sectors (Culwick, 2008). This provides additional
reasons for the research, as well as possible funding into concentrating solar
technologies.
There are also countless institutional benefits that will be gained by the
implementation of CSP technology at the University. This can be expanded to also
include the commercial advantages gained from research at the University.
Research, development and demonstration practices aim at alleviating technical
barriers and reducing costs altogether in improving materials, components and
system design for installers and users. South Africa, because of its traditionally low
cost of coal electricity, has not created an environment where renewable energy is a
viable topic for research and implementation. South Africa has one of the greatest
solar resources in the world and should therefore be technology leaders and
pioneers. With greater emphasis being placed on the need for renewable energy
systems, it is imperative that South Africa develops its skills and a knowledge base
that will work at making the implementation of renewable energy, and in particular
CSP generation, a reality.
According to Sargent and Lundy (2003) there is much R&D still to be done with the
CSP technologies. Countries with the most advanced R&D programs will become the
technology leaders. In the case of renewable energy, the technologies are still
improving and developing while, at the same time, fully market-ready applications
of the technologies are also being continuously improved from experience gained in
commercial applications in the field. Potential R&D efforts should aim to reduce the
cost of mirrors, heliostats, collectors and electric energy generators. Also to develop
and refine thermal energy storage systems that can give up to the critical 12 hours
of thermal storage, which will greatly enhance the economics and potential of solar
thermal electric systems. There is currently research being performed at the
University of Stellenbosch and a number of other universities but this research is
123
mainly focused on Parabolic Trough and Solar Chimney Technology. If the
University of the Witwatersrand were to incorporate a Linear Fresnel System, huge
benefit will be gained locally as well as through the international exposure in CSP
research.
An analysis of different fuels and emission rates is recommended. Ideally if it were

to be incorporated, a renewable fuel such as landfill-gas should be used to take
advantage of the REFIT. However, as mentioned, with an external fuel source there
are certain logistical and fuel storage options that would need to be investigated.
Assuming that the plans for implementation continue, it is advised that insolation
measurements are recorded, at Wits University, or other potential sites. This should
include DNI measurements as well as other meteorological parameters. This ground
measurement will then need to be calibrated with other data sets such as satellite
data. The uncertainty of this data significantly affects the bankability of the project.
The need for full hourly solar resource mapping for the entire country was also
identified. The TMY2 data used in this study (which is only available for
Johannesburg and Cape Town) is not sufficient to make full bankable decisions. This
service should ideally be government sponsored or through the use of institutional
grants, so that the feasibility of such projects is more accessible and uncertainty in
eliminated.
During this study it was found that distributed solar energy can be a possible
solution to the various energy problems faced around the world. A more
appropriate application for distributed solar power generation is possibly in rural
areas where the grid connection costs are high. Many of the existing international
technology providers have the view of developing for large electric capacity’s
(50MW(e) and above). These systems will most likely be locally manufactured but
revenues and intellectual property will remain in hands of international companies.
South Africa in general is blessed with an amazing solar resource yet we are not
124
developing any of our own technology. There is therefore definite scope to develop
niche technology that will break some of the barriers to make this technology
feasible in South Africa. Mr Thomas Roos at the CSIR is currently working on a
heliostat field with an air Brayton cycle turbine. Other technologies that should be
pursued for distributed generation include Linear Fresnel collectors that are easy to
manufacture and don’t involve complicated receiver systems. There is also scope for
developing thermal storage technologies in order to make generation more reliable.
There are also very few off the shelf ORC turbines, especially small scale
(<100kW(e)), which is a technology option worth pursuing.
125
9 CONCLUSIONS AND RECOMMENDATIONS
The objective of the study was to investigate the potential of distributed CSP
integration in urban areas, specifically investigating Johannesburg’s solar resource.
This is done by assessing the technical performance and financial characteristics of
the different technologies in order to identify certain systems that may have the
potential for deployment.
The following conclusions and recommendations address the objectives (Section

1.3) of the report as well and summarise important points identified in the
discussion (Section 8).
9.1 Conclusions
• A relevant literature review was performed which outlines relevant data

with regards the scope of the study.
• Several existing systems, whether they are in research or commercial

operation, were compared.
• Johannesburg has a very intermittent source of DNI solar energy. Even

though the summer months in Johannesburg yield a higher peak DNI, it is
actually the winter months that provide a more consistent average. This is
due to the high amount of cloud cover experienced in summer. The values
obtained for the average yearly insolation was 1781 kWh/m2 based on TMY2
data. With this insolation, CSP electric generation is possible however,
compared to the other locations, it is not ideal. Also, because of its
intermittency is has been advised that certain applications such as HVAC and
process heat and steam requirements be pursued.
126
• A technology screening was performed in order to identify suitable
technologies for application at Wits University. The technical and financial
viability of these technologies identified were then explored. Certain
methodologies and conclusions from this analysis are explored below.
Through the technology screening process, twelve reference systems were

identified for a capacity of 120 kW(e). After a performance and cost analysis for the
reference size plants, three technologies were identified that prove to be the most
suitable for implementation as a distributed energy source for Wits University.
These three technologies are listed below; each with a design capacity of 480 kW(e).
1. CLFR solar collector field with an Organic Rankine Cycle

2. CLFR solar collector field with an Organic Rankine Cycle that integrates
storage for timed dispatch
3. CLFR solar collector field with an Organic Rankine Cycle that integrates
hybridisation with natural gas.
The technologies are intended to be modular in design and would not necessarily be
located on the same site. The field areas and real LECs are summarised in Table 9-1.
With the thermal modelling of the hourly DNI input to the CSP systems, Wits
University’s West Campus Electricity bill was recalculated. The addition of the solar
energy input resulted in certain savings and a new LEC that is Wits-specific. A third
LEC was calculated that integrates an estimated REFIT (R2.05/ kWh). At the time of
writing a CSP REFIT of R2.10/kWh was released and the licensing terms for
independent power production (IPP), using CSP, should be further researched. It is
important to note that the applications considered will not qualify for the REFIT
because the electricity generated is not intended to be sold. It has however been
included in the analysis in order to aid in decision making by indicating what the
market price for this electricity would be. This gives an indication of the price of
electricity generated after this tariff has been taken into account.
127
A summary of the plant areas, LEC as well as the payback of these systems is given
in Table 9-1.
Table 9-1: Summary (480 kW(e) systems)

Average Capacity Factor 0.21 0.32 0.55
Solar Field Area [m2] 6,727 9,923 6,727
Total Plant Area [m2] 10,350 15,266 10,350
Real LEC [R/kWh] 4.31 4.34 3.18
Wits LEC [R/kWh] 3.98 4.00 2.77
Nominal LEC [R/kWh] (with REFIT) 1.93 1.95 2.01
Payback [years] 77.1 72.8 24.6
Payback [years] (with REFIT) 7.8 7.5 6.5
Including hybridisation in the design configurations successfully decreased the peak

usage and resulted in the lowest cost of electricity. It is important to note here that
the hybrid options includes the generation of electricity from natural gas and is not
representative of a solar-only LEC. Hybrid operation will increase emissions and in
urban areas can be a problem. One of the main reasons for using a renewable source
for electricity generation is climate change mitigation and therefore the use of a
renewable fuel should rather be pursued.
The implementation of a storage system results in the highest LEC of the three
options. The use of storage is effective in bringing down the evening peak load
when sunshine is available during day time. Benefits, which will be gained by the
evening dispatch through storage, include energy security and the possibility of a
higher cost saving through the possible introduction of a Time of Use (TOU) charge
but in terms of financial feasibility and space utilisation, energy storage may be less
feasible.
scale CSP systems are higher than conventional fossil fuel options, however several
128
options that may favour implementation were recognised. If the institution values
the CSP generated electricity at the market price as indicated by the CSP REFIT, the
payback time of such systems can be decreased from 73 to 12 years (CLFR, ORC
with storage). Further, due to the scale of the plants analysed, the exploitation of
high efficiencies and economies-of-scale of plants with power levels above 50
MW(e), is not possible. With the introduction of these technologies at lower power
levels, cost savings through the incorporation of other design options (such as using
waste heat for hot water and building HVAC requirements of buildings, process heat
and steam) should be pursued.
9.2 Recommendations
The data comparison of existing technologies is satisfactory but the accuracy of the
results is unknown due to scaling methods used and finding the present value of
past quoted cost data. The economies-of-scale method was used and is suitable for
utility scale plant sizes but discrepancies may arise in the technical performance
when scaling below 1 MW(e). However data was used for a comparative analysis
and thus sufficiently fulfils the scope of work. To gain a full understanding of the
actual cost implications at Wits University, a full cost analysis which would include
equipment as well as implementation costs would need to be performed, for specific
designs. The work scope for this study, as mentioned, did not include independent
research for information or developing independent cost-estimates.
The analysis excluded conventional diesel generators and other distributed

renewable technologies such as photovoltaic systems. A comparison of these
technologies to the solar systems investigated would be useful in showing the cost
of distributed power options as opposed to grid-connected power and is
recommended for future investigations.
The design of these three configurations was based on common annual capacity
factors. This method, was chosen because the same electric production between
129
plants was necessary for level comparison. It is recommended that the three chosen
configurations be further designed according to the design point peak DNI. This
design procedure should also be cost-optimised.
Price parity will be reached in the near future where the cost of CSP technology will
be more competitive than traditional fossil-fuelled power generation. It is
recommended that financial predictions be performed with respect to the South
African electricity tariffs (as well as other distributed sources of power) to aid
investor decision-making.
Carbon financing and the sale of TRECS have not been incorporated into the analysis
and it is recommended that these be further investigated. These should be
incorporated into a full discounted cash flow to predict the future costs of
electricity. Other factors to be considered is carbon taxing that, if implemented in
South Africa, would influence the price of electricity.
South Africa has one of the greatest solar resources in the world and should
therefore be technology leaders and pioneers. With greater emphasis being placed
on the need for renewable energy systems, it is imperative that South Africa
develops its skills and a knowledge base that will work at making the
implementation of renewable energy, and in particular CSP generation, a reality.
There are countless institutional benefits that will be gained by the implementation
of CSP technology at the University. This can be expanded to also include the
commercial advantages gained from research at the University. Research,
development and demonstration practices aim at alleviating technical barriers and
reducing costs altogether in improving materials, components and system design
for installers and users.
Technologies identified that should be pursued for distributed generation include

Linear Fresnel collectors that are easy to manufacture and don’t involve
complicated receiver systems. There is also scope for developing thermal storage
130
technologies in order to make generation more reliable. There are also very few off
the shelf ORC turbines, especially small scale (<100kW(e)), which is a technology
option worth pursuing.
Because of the lack of solar resource data in the country, it is advised that insolation
levels are measured. This is necessary if plans for implementation continue. This
ground measurement will then need to be calibrated with other data sets such as
satellite data.
131
REFERENCES
Aitken D (2003) Transitioning to a Renewable Energy Future - White Paper,

International Solar Energy Society White Paper 2003
Becker M, Meinecke, Geyer M, Trieb F, Blanco M, Romero M, Ferrière A (2000) Solar

Thermal Power Plants, Prepared for the EUREC-Agency
Beerbauma S, Weinrebeb G (2000) Solar thermal power generation in India-a

techno-economic analysis, Renewable Energy 21 pp.153-174
Black and Veatch Corporation (2007) Arizona Renewable Energy Assessment,

Arizona Public Service Company, Salt River Project, Tucson Electric Power
Corporation
Bureau of Land Management (BLM) (2003) Assessing the Potential for Renewable
Energy on Public Lands, U.S. Department of the Interior Bureau of Land Management
DOE/GO-102003-1704
Brink E (2008) Matlab © source code for Wits University bill adjustments
Broesamle H, Mannstein H, Schillings C, Trieb F (2000) Assessment of Solar

Electricity Potentials in North Africa Based On Satellite Data and a Geographic
Information System, Prepared for DLR
CDM UNFCC (2009) Clean Development Mechanism, INTERNET, Last Accessed

02/02/2009, http://cdm.unfccc.int/about/index.html
City Of Johannesburg (2008) Schedule of Tariffs for 2008/2009
132
CSP (2008) Concentrating Solar Power, INTERNET - http://www.solar-energy-at-
home.com/concentrating-solar-power.html
Culwick L (2008) An Investigation into the Solar Resource for Electricity Generation
in South Africa, Final year research project reports of the School of Mechanical,
Industrial and Aeronautical Engineering and the Built Environment, University of
the Witwatersrand, Johannesburg
Department of Minerals and Energy (DME) (2003) White paper on Renewable

Energy, DME, South Africa
Department of Minerals and Energy (DME) (2009) Tradable Renewable Energy

Certificates (TRECS) – INTERNET, Last Accessed 04/02/2009,
http://www.dme.gov.za/energy/renew_TRECS.stm
Egoli Gas (EG) (2009) INTERNET, Last Accessed 05/11/2008,

http://www.egoligas.co.za/
Engineering Toolbox (2009) Boiler room Sizes, INTERNET - Last Accessed

05/11/2008, http://www.engineeringtoolbox.com
Eskom (2008) Annual Report, Corporate governance and tables - INTERNET - Last
Accessed 02/02/2009 –
http://www.eskom.co.za/annreport08/ar_2008/table_one_kilowatt.htm
Eskom (2009) Power Purchase Programmes; An initiative by Eskom to bridge the

Gap between the supply and demand of energy, INTERNET - Last Accessed
02/02/2009 – http://www.eskom.co.za/live/content.php?Item_ID=6865, lat
accessed
133
Eskom (b) (2009) New build Programme, INTERNET - Last Accessed 02/02/2009 –
http://www.eskom.co.za/live/content.php?Item_ID=5981
Fin24 (2009) “Green Energy Tariffs Hiked”, INTERNET -Last Accessed 21/03/2009–
http://www.fin24.com/articles/default/display_article.aspx?Nav=ns&ArticleID=15
18-25_2494501
Freepower (2008) Renewable Energy and Energy Saving Solutions, INTERNET - Last
Accessed 09/11/2008 – http://www.freepower.co.uk
Geyer M (1999) Concrete Thermal Energy Storage for Parabolic Trough Plants-A
Proposal to the 5th Framework Program of the European Union, Deutsches Zentrum
für Luft- und Raumfahrt, Plataforma Solar de Almeria, Spain
Geyer M (2007) Introducing Concentrated Solar Power on the International Markets

Worldwide Incentives, Policies and Benefits, Solar Paces/Abengoa Solar
Haaf W, Friedrich K, Mayr G, Schlaich J (1983) Solar Chimneys Part 1: Principle and
Construction of the Pilot Plant in Manzanares. International Journal of Solar Energy
2, pp. 3–20
Häberle A , Zahler C, Lerchenmüller H (2002) The Solarmundo line focussing Fresnel

collector Optical and thermal performance and cost calculations, Fraunhofer Institute
for Solar Energy Systems
Hassani V, Price H (2001) Modular Trough Power Plants, Proceedings of Solar Forum
2001 Solar Energy, Washington, DC, pp.1-7, © ASME
Hosseini R, Soltani M, Valizadeh G (2005) Technical and economic assessment of the

integrated solar combined cycle power plants in Iran, Renewable Energy 30 1541–
1555
134
Hoff T (2000) The Benefits of Distributed Resources to Local Governments: An
Introduction, Clean Power Research, Napa, CA 94558
Huber (2008) Value Engineering and Analysis, VM Services handbook lecture series
2008
IEEE (2005) Standard for Interconnecting Distributed Resources with Electric Power
Systems, IEEE 1547
International Emissions Trading Association (IETA) (2009) Europe Fails Kyoto

Standards as Trading Scheme Helps Polluters, INTERNET - Last Accessed
10/03/2009 –http://www.ieta.org/ieta/www/pages/index.php?IdSitePage=1149
International Energy Agency (IEA) (1991) Guidelines for the Economic Analysis of
Renewable Energy Applications, IEA, Quebec
ISRAEL21c (2007) Israel's Solel to build largest solar park in world in California,
INTERNET - Last Accessed 16/10/2008 –
http://www.israel21c.org/bin/en.jsp?enDispWho=Articles^l1721&enPage=BlankP
age&enDisplay=view&enDispWhat=object&enVersion=0&enZone=Technology
Kolb G (1998) The Economic Evaluation of Solar-only and Hybrid power Towers using
Molten Salt Technology, Solar Energy Vol 62 No1
Lotker, M. (1991) Barriers to Commercialization of Large-Scale Solar Electricity:

Lessons Learned from the LUZ Experience, Sandia National Laboratories,
Albuquerque, New Mexico: Report SAND91-7014.
Matlab © (2005) Version 7, The Mathworks Inc.
135
MOL (2008) Metoring online Muncipal Electricity Bills, INTERNET - Last Accessed
28/02/2009 –http://www.meteringonline.org.za
Mills D, Morrison G, Lievre P (2003) Project Proposal for a Compact Linear Fresnel
Reflector Solar in Thermal Plant in the Hunter Valley, School of Physics,
University of Sydney, Australia
Mills D, Morrison G (2000) Compact Linear Fresnel Reflector Solar Thermal

Powerplants, Solar Energy, Vol 68 No3, pp.263-283
National Renewable Energy Laboratory (NREL) (1995) User's Manual for TMY2s
Meteorological Years Derived from the 1961-1990 National Solar Radiation Data
Base, CD-ROM. Golden, Colorado, NREL/SP-463-7668
National Renewable Energy Laboratory (NREL) (2008) DNI Radiation Data,

INTERNET - Last Accessed 14/10/2008 –
http://www.nrel.gov/csp/troughnet/pdfs/2007/stoffel_solar_data.pdf
NERSA (2008) Renewable Energy Feed - In Tariff – December, 2008, NERSA

Consultation Paper, South Africa
Obernberger I, Thornhofer P, Reisenhofer E, (2002) Description and Evaluation of

the New 1000kW(e) Organic Rankine Cycle Process Integrated in the biomass CHP
plant, Lienz, Austria, Euroheat and Power, Vol 10/1002
Peters M, Timmerhaus E (1991) Plant Design and Economics for Chemical Engineer,
McGraw-Hill, pp.164-266
Philibert C (2005) The Present And Future Use Of Solar Thermal Energy As A Primary
Source Of Energy, International Energy Agency, Paris, France
136
Pilkington Solar International GmbH (2000) Survey of Thermal Storage for Parabolic
Trough Power Plants, NREL Technical Report, Subcontract No. AAR-9-29442-05
Pitz-Paal R, Dersch J, Milow B (2005) European Concentrated Solar Thermal Road-

Mapping (ECOSTAR), Road Map Document (WP 3 Deliverable Nº 7)
Power Technology (2009) Solar Arguments Heat Up, INTERNET - Last Accessed
21/01/2000 – http://www.power-technology.com/features/feature1480/
Price H, Kearney D (1999) Parabolic-Trough Technology Roadmap: A Pathway for

Sustained Commercial Development and Deployment of Parabolic- Trough
Technology, U.S. Department of Energy (DOE)
Price H, Kearney D (2003) Reducing the Cost of Energy from Parabolic Trough Solar
Power Plants Preprint, National Renewable Energy Laboratory
Prabhu E (2006) Solar Trough Organic Rankine Electricity System (STORES) Stage 1:
Power Plant Optimization and Economics, Subcontract Report, NREL/SR-550-39433
Roos T (2009) Personal Communication, DPSS, CSIR
Sargent and Lundy (S&L) (2003) Assessment of Parabolic Trough and Power Tower
Solar Technology Cost and Performance Forecasts, Subcontractor Report, NREL/SR-
550-34440,
Schaffler (2007) Tradable Renewable Energy Certificate System Feasibility,

Prepared for the Department of New and Renewable Energy Pretoria, Report
Number DME/CE/001/260301, pp 8
Sitenet (2008) Solar Power Systems, INTERNET - Last Accessed 18/10/2008 –

http://www.sitenet.com/futureSeries/2008/Solar-Power.htm
137
Solargenix Energy (2005) Solar Thermal Parabolic Trough Electric Power Plants for
Electric Utilities in California, California Energy Commission, Public Interest Energy
Research (Pier) Program, CEC-500-2005-175
Solar Thermal Industry Association (STI) (2005) Concentrating Solar Thermal Power
– NOW!, Solar Paces, Estia, Greenpeace
Solar Heat and Power Europe (2008) Sustainable Power, INTERNET - Last Accessed
15/11/2009 – http://www.shp-europe.com/
Stine W, Geyer M (2008) Power from the sun, INTERNET - Last Accessed
10/12/2009 –http://www.powerfromthesun.net
Standard Bank (SB) (2008) Standard Bank Fees and rates, INTERNET - Last Accessed
10/11/2009 –
http://www.standardbank.co.za/SBIC/Frontdoor_02_05/0,2358,3447_1091713_0,0
0.html
Stoddard L, Abiecunas J, O'Connell R (2006) Economic, Energy, and Environmental

Benefits of Concentrating Solar Power in California, Subcontract Report NREL/SR-
550-39291
Trieb F, Langnib O, Klaib H (1997) Solar electricity Generation-A Comparative view of

Technologies, Costs, and Environmental Impact, Solar Energy, Vol 59, Nos. 1-3, pp.89-
99
UNFCC (1998) Kyoto Protocol to the United Nations Framework Convention on

Climate Change, Article 12, UNFCCC pp. 11.
138
Van Heerden (2001) Research Report, CSP Pre-feasibility Study, Eskom Research
RES/RR/01/15662
Westney (1997) The Engineer’s Cost Handbook, Marcal Dekker, New York, USA
World Bank (2006) Assessment Of The World Bank Group/Gef Strategy For The
Market Development Of Concentrating Solar Thermal Power, The International Bank
for Reconstruction and Development, Washington D.C., USA
Wolf C Jnr (1969) The Present Value of the Past, The Rand Corporation, P4067
XE (2008) Exchange rates, INTERNET Last Accessed 11/11/2009 –

http://www.xe.com
139
APPENDIX A STORAGE CONCEPTS
Storage concepts can be classified as active or passive systems. Active storage is

mainly characterised by forced convection heat transfer into the storage material.
The storage medium itself circulates through a heat exchanger. This heat exchanger
can also be a solar receiver or a steam generator.
The main characteristic of a passive system is that a heat transfer medium passes
through storage only for charging and discharging. The heat transfer medium itself
does not circulate.
Active Thermal Energy Storage

Active thermal systems typically utilize tank storage. They can be designed as one
tank or two tank systems. Active storage is again subdivided into direct and indirect
systems. In a direct system the heat transfer fluid, which collects the solar heat,
serves also as the storage medium, while in an indirect system, a second medium is
used for storing the heat. An example of the two-tank systems for solar electric
applications are the storage systems of the SEGS I (Pilkington, 2000). Figure 1
shows a schematic flow diagram of SEGS I.
Figure A1: Schematic Flow Diagram of the SEGS 1 plant
140
A two-tank system uses one tank for cold heat transfer fluid (HTF) coming from the
steam generator and one tank for the hot HTF coming directly out of the solar
receiver before it is fed to the steam generator. The advantage of this system is that
cold and hot HTF are stored separately. The main disadvantage is the need for a
second tank (Pilkington, 2000).
The single-tank system reduces storage volume and cost by eliminating a second
tank. However, in a single-tank system it is more difficult to separate the hot and
cold HTF. Because of the density difference between hot and cold fluid, the HTF
naturally stratifies in the tank, from coolest layers at the bottom to warmest layers
at the top. These systems are called thermocline storage. Maintaining the thermal
stratification requires a controlled charging and discharging procedure, and
appropriate methods or devices to avoid mixing. Filling the storage tank with a
second solid storage material (rock, iron, sand etc.) can help to achieve the
stratification (Pilkington, 2000).
Passive Thermal Energy Storage

Passive systems are generally dual medium storage systems. The HTF carries
energy received from the energy source to the storage medium during charging and
receives energy from the storage material when discharging. These systems are also
called regenerators.
The storage medium can be a solid, liquid, or phase change medium. In general, a
chemical storage system employs at least two media.
The main disadvantage of regenerators is that the HTF temperature decreases

during discharging as the storage material cools down. Another problem is the
internal heat transfer. Especially for solid materials, the heat transfer is rather low,
and there is usually no direct contact between the HTF and the storage material as
the heat is transferred via a heat exchanger (Pilkington, 2000).
141
APPENDIX B TECHNOLOGIES USED IN THE COMPARISON
The majority of the technologies that have been used in this comparison have been
specified in the Ecostar study (2003) (where information is not sourced from the
Ecostar Study it is appropriately referenced). This is done to keep the information as
closely comparable as possible. This section serves as a basic reference to the
background of each technology.
Parabolic Trough with Storage

Ecostar based their model on a few existing technologies. These include all the SEGS
plants in the USA as well as two 50 MW plants built in Guadix in the province of
Granada/Spain. Based on these reference data, they designed a power plant for the
selected site, load curve and other boundary conditions with the lowest solar LEC
according to the model. The degree of detail in their model appears sufficient to
analyse the overall impact of changes in cost and performance.
Parabolic Trough with Direct Steam Generation (DSG)

The need for high operating temperatures forced the developer of the existing SEGS
plants, the company LUZ, to work in the solar field with thermal fluids (synthetic
oils) able to withstand 400ºC. One of the most important objectives of LUZ was the
replacement of this expensive heat carrier by water which is directly heated up and
converted into superheated steam in the absorber pipes of parabolic trough
collectors. During the first phase of the EU co-funded DISS project (1996-1998) a
life-size solar test facility was designed and implemented at the Plataforma Solar de
Almería (PSA) to investigate under real solar conditions the DSG process and
evaluate the open technical questions concerning this new technology.
Once the feasibility of the DSG process was proven in the project DISS 10, and
design/simulation tools had been developed, the EU co-funded project INDITEP
(2002-2005) undertook the detail design of a first pre-commercial DSG power plant
of 5 MW(e). The optimisation of some key components for DSG plants (e.g.,
142
water/steam separators, selective coatings, etc.) was also included in the work
program of INDITEP.
The reference plant used in the Ecostar evaluation is composed of ten INDITEP
plants, working in parallel, with a net electrical power of the DSG reference plant
being 47 MW(e). No storage system is foreseen because of technical problems.
CLFR
The linear Fresnel system may be considered as innovation for the direct steam
generating (DSG) parabolic trough system, since it is also designed for DSG rather
than for the utilisation of a heat transfer fluid. Since the plant design and
characteristics differ significantly from a parabolic trough plant, Linear Fresnel
systems have been treated in a separate model and the results are shown in a
special manner compared to the other innovations. Linear Fresnel systems suffer
from performance drawbacks due to higher intrinsic optical losses compared to
parabolic trough systems. The model is a 50 MW system that is based on the
performance data given by the Fraunhofer Institute.
Central receiver with Molten Salt

Several molten salt development and demonstration experiments have been
conducted over the past two and half decades in the USA and Europe to test entire
systems and develop components. The largest demonstration of a molten salt
central receiver was the Solar Two project- a 10 MW central receiver located near
Barstow, CA.
The Solar Tres concept is considered as the current state of the art for molten salt
central receivers. Thus, a 50 MW reference system composed of several modules
based on 17 MW Solar Tres project with molten salt technology has been sized to
accomplish with the common restriction agreed in this project.
143
Central receiver with Atmospheric air
A central receiver solar power plant with an atmospheric air heat transfer circuit
based on the so-called PHOEBUS scheme, where atmospheric air is heated up
through a porous absorber receiver to temperatures in the order of 700ºC and used
to produce steam at 480-540ºC and 35-140 bar.
This concept has been studied by the German company TSA and the operational
results attracted the interest of the Spanish company Abengoa that decided to
analyze the Phoebus scheme as one of the options for the design of its first
commercial demonstration plant. The project named PS10 started in 1999 and its
goal is the construction and connection to the grid of a 10 MW plant is located in
Seville (Spain).
Central receiver using pressurised air in combination with a solar hybrid gas-
turbine
This is based on the Refos receiver type, which is a pressurised volumetric air
receiver. Differently from all other concepts, solar high temperature heat is
introduced into a gas-turbine. The concept needs additional fuel to increase the
temperature above the level achieved by the solar system.
This concept has been investigated in a project with the following partners: ORMAT,
(Israel) CIEMAT (Spain), DLR (Germany), SOLUCAR (Spain) and TUMA
(Switzerland). This project has included experimental investigations of a REFOS
system at the Plataforma Solar at Almería, Spain as well as theoretical studies
concerning the up scaling of the plant to 16 MW(e). In the experimental part of the
SOLGATE project a cluster of three receivers with 1 MWth power was integrated
into a gas turbine with a design power output of 250 kW.
Since the reference system (Solgate PGT10) only has a capacity of 14.6 MW(e), a
power park of four equal systems at one site is investigated to account for similar
O&M conditions. Specific costs for the power block, receiver, and storage were
144
scaled using an exponent of 0.93 resulting in 90% of the specific costs figures of the
original design.
Dish Stirling
Seven 10 kW EURODISH systems are currently in operation in several countries
(Spain, Italy, France, Germany, India). A WGA dish with a SOLO Stirling engine is still
running at the Sandia National laboratory. Numerous solar receivers were also
designed and tried with the Stirling engines. More recent designs (e.g. the SAIC
system) include a fuel combustion option to boost power during periods of
insufficient solar input.
Based on these reference data a power plant for the selected site was designed.
Since one reference system unit has only a power capacity of 22 kW(e) a power park
of many equal systems at one site, to account for similar O&M conditions, was
investigated.
ISCCS
During the late nineties and earlier this century, the Global Environment Fund (GEF)
and the World Bank considered a number of ISCCS configurations. Spencer
Management Associates found the incremental solar costs for a 30 MW power plant
to be below $0.1/kWh. Eskom took this study and detailed their findings for a site in
Upington.
Solar Chimney
In 1979, a prototype power plant employing the solar chimney concept was funded
by the Federal German Ministry of Research and technology. A site was provided by
the Spanish utility Union Electrica Fenosa in Spain. The utility completed
construction in 1982 with a peak design output of 50 kW. Schlaich, Bergermann and
Partner designed three further plants at 5, 30 and 100 MW (Schlaich et al., 1995).
The chimneys were designed for operation in Manzanares which receives a Global
145
insulation of 2301 kWh/m2/a. The Eskom study adjusts these values to values
expected in South Africa.
Modular Parabolic Trough plants using Organic Rankine Cycles (ORC)

A preliminary analysis has been completed to assess the potential economic
feasibility of small trough ORC power plants. NREL has developed an hourly
simulation model capable of modelling the performance of parabolic trough solar
power plants. Using the ORC power cycle performance for the system developed by
Barber Nichols, NREL has modified the trough power plant model to predict the
performance from a parabolic trough ORC plant. A nominally 1 MW(e) net parabolic
trough ORC power plant with thermal storage was modelled for this analysis
(Hassani, 2001). The assumptions used in the model are listed in the paper but some
important differences in their model are given here. The location of the analysis is
Barstow California, which has an annual DNI of 2800 kWht/m2. The solar field
availability was assumed to be 99% as opposed to 96% used in the Ecostar study. It
is based on thermal storage of 9 hours. The financial data such as the discount rate
and annual insurance, gives a FCR of 12.25%. (In the comparison and analysis, this
value has been adjusted down to Ecostar level 9.88%).
146
Table B1 Summary of Evaluated Technologies
Property SEGS SEGS with Storage SEGS DSG
Collector Parabolic Trough Parabolic Trough Parabolic Trough

Receiver Linear receiver (tubes) Linear receiver (tubes) Linear receiver (tubes)
2
Total Area of Plant [m ] 884070 884070 884070
Cycle Rankine Steam Rankine Steam Rankine Steam
Storage None Two-tank molten Salt Storage No Storage System
Storage Capacity [h] 0 3 0
This is the reference case, several 30 MW 50 MW Andasol I completed 2008 and Andasol II, 4.7 MW INDITEP study - DLR at PSA in
Planned/built power size
hybrid plants are currently operating in the US under preparation, Spain Spain
Several 80MW plants built in US operating Several 80MW plants built in US operating since
Maturity Single row experimental plant in Spain
since the 1980's the 1980's
Temperature 390 393 411

Reference Size 100 50 10x4.7
Solar Capacity Factor 25% 29% 22%
147
Property CLFR ISCC Central receiver Molten Salt Central receiver Atmospheric Air
Collector CLFR Parabolic Trough Heliostat field Heliostat Field

Fixed Linear
Receiver Linear receiver (tubes) Molten Salt Receiver Saturated Steam Central Receiver
Receiver
Total Area of
2 752400 916320 1045800
Plant [m ]
Cycle Rankine Steam Combined Cycle Rankine Steam Rankine Steam
Storage None 2-tank-molten-salt storage Water/steam buffer storage
Storage
0 3 0.4
Capacity [h]
Planned/built
World bank Studies Solar Tres (17MW), planed, Spain PS 10 (11MW),
power size
Solarmundo Solar 2 (11 MW) experimental plant in Several experimental plants up to

Maturity Algeria 140 MW ISCCS
prototype California in the 1990s 2MWth have been tested
Temperature 411 400 565 680

Reference
50 50 3x17 5x11
Size
Solar
Capacity 18% 25% 33% 26%
Factor
148
Property Central receiver Brayton Solar Chimney Dish Stirling MTPP
Collector Heliostat Field Solar Collector Parabolic Dish Parabolic Trough

Receiver Chimney and Turbine Cavity receiver with tube bundle Linear Receiver
Total Area of Plant
2 942000 700000
[m ]
Cycle Combined Cycle Stirling Cycle Organic Rankine Cycle
Two-tank thermal
Storage No storage system available None to Date
storage system
Storage Capacity [h] 0 0
Planned/built power
Solgate study 14.6 MW(e) 22kW
size
About 30 units up to 25 kW(e) are in

Maturity Pilot Plant in Australia
operation at different sites
Temperature 800 800 304

Reference Size 4x14.6 100 2907 × 25 kW(e) 1 x 50MW
Solar Capacity
19% Solar Only (55%) 33% 22% 54%
Factor
149
APPENDIX C SOLAR RADIATION DATA
Table C1: Average Hourly Statistics for Direct Normal Solar Radiation Wh/m²
Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec
1 0:01- 1:00 0 0 0 0 0 0 0 0 0 0 0 0
2 1:01- 2:00 0 0 0 0 0 0 0 0 0 0 0 0
3 2:01- 3:00 0 0 0 0 0 0 0 0 0 0 0 0
4 3:01- 4:00 0 0 0 0 0 0 0 0 0 0 0 0
5 4:01- 5:00 0 0 0 0 0 0 0 0 0 0 0 0
6 5:01- 6:00 0 0 0 0 0 0 0 0 0 0 0 0
7 6:01- 7:00 88 54 12 0 0 0 0 0 12 76 142 151
8 7:01- 8:00 250 233 289 234 263 203 158 177 246 271 262 296
9 8:01- 9:00 356 366 446 467 525 486 477 472 479 436 386 433
10 9:01-10:00 411 401 508 591 682 671 632 634 603 530 424 484
11 10:01-11:00 416 405 523 624 740 736 729 726 657 576 454 499
12 11:01-12:00 385 423 498 591 759 769 799 791 643 551 425 467
13 12:01-13:00 361 394 482 528 747 772 821 785 640 491 413 424
14 13:01-14:00 360 389 410 483 715 762 823 769 623 454 352 399
15 14:01-15:00 358 378 426 478 685 719 781 731 590 403 312 343
16 15:01-16:00 323 352 396 409 589 628 675 611 499 330 263 288
17 16:01-17:00 253 289 341 263 288 361 345 314 304 184 177 210
18 17:01-18:00 145 177 158 11 0 0 0 0 23 24 43 104
19 18:01-19:00 2 3 0 0 0 0 0 0 0 0 0 0
20 19:01-20:00 0 0 0 0 0 0 0 0 0 0 0 0
21 20:01-21:00 0 0 0 0 0 0 0 0 0 0 0 0
22 21:01-22:00 0 0 0 0 0 0 0 0 0 0 0 0
23 22:01-23:00 0 0 0 0 0 0 0 0 0 0 0 0
24 23:01-24:00 0 0 0 0 0 0 0 0 0 0 0 0
Sum Month [Wh/m2]114948 108192 139159 140370 185783 183210 193440 186310 159570 134106 109590 127038
Max Hour 11 12 11 11 12 13 14 12 11 11 11 11
Min Hour 1 1 1 1 1 1 1 1 1 1 1 1
Sum Year [kWh/m2a]1781.716
150
Figure C1: Average Daily Data - JHB
Figure C2: Monthly Statistics-JHB
151
Hourly DNI Data for June - Johannesburg
1000
900
800
700
600
DNI [kW/m2]
500
400
300
200
100
0
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23
Time
Figure C3: Hourly DNI Data - June- JHB
Hourly DNI Data for January - Johannesburg
1200
1000
800
DNI [kW/m2]
600
400
200
0
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23
Time
Figure C4: Hourly DNI Data - January- JHB

152
APPENDIX D DATA VERIFICATION
The optical efficiency obtained is 76 % and is confirmed by both NREL and

Broesamle et al..
To verify the Ecostar results a design net electric output of 50MW is used. By using a
design peak DNI of 950 W/m2 and equation (16) results in an aperture area of 451
488 m2, which includes three hours of storage and a solar multiple of 1.4. The value
for the solar field in the Ecostar study is 442035 m2. The calculated area is 1.9%
larger (due to lower solar field efficiency) which is a satisfactory result and confirms
the model for the parabolic trough. For the CLFR with no storage in the Ecostar
design, the method was again verified with a slightly lower difference of 0.4%.
The details are summarised in Table D1.
Table D1: Parabolic Trough and CLFR Verification
Parabolic Trough CLFR

Ecostar Calculated Ecostar Calculated
Pnet 50MW 50MW 50MW 50MW
DNI 2,014 2,014 2,014 2,014
nopt 75% 76% 64% 68%
nsf 54.2% 53.8% 42.2% 42.0%
ns-e 14.08% 13.97% 10.54% 10.51%
Aa 442035 451,488 376200 377,567
Enet 124,670,470 127,020,000 79,886,327 79,891,200
Difference 1.9% 0.4%
153
APPENDIX E JOHANNESBURG ELECTRICITY RATES
Wits University is medium voltage kVA business customer and the following charge
rates are applicable (City of Johannesburg, 2008). A 2% surcharge will be levied on
business and large power users.
Service Charge: R 1194.14 per month
Energy Charge (Subject to a seasonal change):

The summer rate is September through to April with both months inclusive (8
Months) and the winter rate is May through to August (4 Months)
Summer: 23.45 cents per kWh

Winter: 34.68 cents per kWh
Demand Charge:
R 78.24 per kVA
R 80.67 per kVA
Reactive Energy Charge:

6.13 cents per kVARh supplied in excess of 30% (0,96PF) of kWh recorded during
the entire billing period. The excess reactive energy is determined using the billing
period total.
Minimum Demand Charge Determination

The minimum demand charge payable monthly in terms of this tariff shall be
calculated using the greater of:
(i) The measured demand;
(ii) A demand of 70 kVA
(iii) A demand based on the 80% average of the three highest demands recorded
over the preceding 12 months.
154
APPENDIX F NATURAL GAS PRICING TARIFF FROM EGOLI
GAS
Properties
Energy Content: 36.10MJ/m3 (@15C and 101.3 kPa)
TARIFF BANDS:
Tariff Band 01 - R 158.19 (Excl. VAT)
0 to 599 GJ/Annum

600 to 2,399 GJ/Annum

2,400 to 4,799 GJ/Annum

4,800 to 9,599 GJ/Annum

9,600 to 17,999 GJ/Annum

18,000 to 35,999 GJ/Annum

36,000 to 119,999 GJ/Annum
MONTHLY BASIC CHARGES:

R 242.36 (Excl. VAT)
There is an annual tariff increase on the 1st of July.
155
APPENDIX G WITS UNIVERSITY USAGE AND BILLING
TRENDS
West Campus Usage November 2007
2500
2000
Electricity Usage [kW]
1500
1000
500
0
0 5 10 15 20
Time (hour)
Figure G1: West Campus Usage- November 2007
West Campus Usage June 2008
3000
2500
2000
Electricity Usage [kW]
1500
1000
500
0
0 5 10 15 20
Time (hour)
Figure G2: West Campus Usage- June 2008
156
Figure G3: Monthly bill June 2008
Figure G4: Historical Billing Trend for West Campus
157
APPENDIX H SPECIFIC COSTS
Table H1: Specific Costs
Source Ecostar MTPP Eskom Solar mundo BIO ORC

Parabolic Parabolic Parabolic Parabolic Parabolic Parabolic Parabolic Parabolic
CSP System CLFR CLFR CLFR CLFR bio bio
Trough Trough Trough Trough Trough Trough Trough Trough
Capacity 50MW 120kW 50MW 120kW 1MW 120kW 100MW 120kW 50MW 120kW 50MW 120kW 1MW 120kW
Specific Investment cost for
€ 251.59 € 341.84 € 146.56 € 199.13 € 196.66 € 219.03 € 291.65 € 541.98 € 280.78 € 429.08 € 117.00 € 129.35 € 0.00 € 0.00
Solar Field (*/m2)
Specific Investment cost for
€ 854.91 € 1,306.60 € 854.91 € 1,306.60 € 1,671.59 € 1,940.35 € 733.00 € 1,431.95 € 933.56 € 1,198.60 € 933.56 € 1,198.60 € 1,809.43 € 2,100.27
Power Block [*/kW(e)l]
Specific Land Cost [*/m2] € 2.44 € 2.44 € 2.44 € 2.44 € 1.44 € 1.44 € 0.00 € 0.00
Specific Investment in
€ 37.86 € 57.21 € 17.31 € 20.01 € 32.73 € 62.89 € 0.00
Storage[*/kWhth]
Surcharge for Construction,
Engineering and 20.00% 20.00% 20.00% 20.00% 10.00% 10.00% 20.00% 20.00% 10.00% 10.00% 10.00% 10.00% 0.00% 0.00%
Contingencies %
Average O&M costs [% of total
2.50% 3.75% 2.50% 3.75% 1.50% 3.00% 0.00% 0.00%
capital costs]
158
APPENDIX I PERSPECTIVE MODEL
Table I1: Perspective Model for Twelve Alternatives
Functions
3. Simplify integration
1. Produce Electricity
4. Reduce Emissions
2. Minimise costs
Rank
Total
Alternatives
Score 13 11 9 9
1 PT 5 65 5 55 7 63 10 90 273 10
2 PT, Stor 7 91 5 55 5 45 10 90 281 9
3 PT, Hybrid 10 130 10 110 7 63 5 45 348 4
4 PT, ORC 5 65 4 44 5 45 10 90 244 12
5 PT, Stor, ORC 7 91 4 44 4 36 10 90 261 11
6 PT, Hybrid, ORC 10 130 9 99 5 45 5 45 319 6
7 CLFR 5 65 7 77 10 90 10 90 322 5
8 CLFR, Stor 7 91 7 77 9 81 10 90 339 3
9 CLFR, Hybrid 10 130 10 110 10 90 5 45 375 1
10 CLFR, ORC 5 65 6 66 9 81 10 90 302 7
11 CLFR, Stor, ORC 7 91 6 66 6 54 10 90 301 8
12 CLFR, Hybrid, ORC 10 130 10 110 9 81 5 45 366 2
159
APPENDIX J REFERENCE PLANT RESULTS
Table J1: Reference Plant Results
CLFR,
PT, PT, Stor, PT, Hybrid, CLFR, CLFR, Stor,
PT PT, Stor PT, ORC CLFR CLFR, Stor CLFR, ORC Hybrid,
Hybrid ORC ORC Hybrid ORC
ORC
Total Electricity Consumption [kWh] 12,922,073 12,922,073 12,922,073 12,922,073 12,922,073 12,922,073 12,922,073 12,922,073 12,922,073 12,922,073 12,922,073 12,922,073
Total Solar Electricity Generated [kWh] 206,424 312,376 569,400 211,466 319,725 569,400 209,641 317,484 569,400 214,714 324,954 569,400
Yearly Bill [R] 7,291,654 7,291,654 7,291,654 7,291,654 7,291,654 7,291,654 7,291,654 7,291,654 7,291,654 7,291,654 7,291,654 7,291,654
Total Bill [R] (incl. cost of Solar) 8,182,181 8,564,679 8,598,830 8,453,947 9,045,429 9,162,401 7,951,079 8,220,164 8,353,900 8,107,095 8,531,930 8,803,593
Extra Cost for Solar [R/year] 890,527 1,273,025 1,307,176 1,162,293 1,753,775 1,870,747 659,425 928,510 1,062,246 815,441 1,240,275 1,511,939
Cost Saved on Bill [R/year] 106,501 163,905 298,532 108,615 167,774 298,667 107,863 166,810 298,620 109,976 170,023 298,753
Solar LEC [R/kWh] 4.83 4.60 2.82 6.01 6.01 3.81 3.66 3.45 2.39 4.31 4.34 3.18
Real Solar LEC [R/kWh] 4.41 4.17 2.37 5.59 5.58 3.36 3.24 3.02 1.94 3.89 3.91 2.73
Average Capacity Factor 0.20 0.30 0.55 0.20 0.31 0.55 0.20 0.31 0.55 0.21 0.31 0.55
Total Investment [R] 7,015,875 10,020,564 7,015,875 9,281,377 13,919,049 9,281,377 5,317,550 7,522,350 5,317,550 6,660,049 10,063,111 6,660,049
Payback [years] 65.9 61.1 23.5 85.5 83.0 31.1 49.3 45.1 17.8 60.6 59.2 22.3
- - - - - - - - - - - -
Nominal LEC [R/kWh] (with REFIT) 2.36 2.12 1.61 3.54 2.01 2.04 1.19 0.97 1.18 1.84 1.86 1.97
Extra cost for Solar [R/year] (with REFIT) 467,358 632,654 874,853 728,788 612,356 1,117,873 229,660 277,668 629,923 375,278 574,121 1,079,616
Payback [years] (with REFIT) 13.2 12.5 9.6 17.1 10.6 8.8 9.9 9.2 7.3 12.1 12.0 9.1
160
LEC
Solar LEC
7.00 Real LEC
6.01 6.01 Nominal LEC (with FIT)
6.00 5.59 5.58
4.83
5.00 4.60
4.41 4.31 4.34
4.17
3.81 3.89 3.91
4.00 3.66
LEC [R]
3.54 3.45
3.36 3.24 3.18
3.02
3.00 2.82 2.73
2.36 2.37 2.39
2.12 2.01 2.04 1.94 1.84 1.86 1.97
2.00 1.61
1.19 1.18
0.97
1.00
-
Hybrid ORC Hybrid, Stor Hybrid ORC Stor, ORC Hybrid,
ORC ORC
Technology Option
Figure J1: LEC for 120 kW(e) Reference Plants
Payback Payback
90.0 85.5 Payback - with FIT
83.0
80.0
70.0 65.9
61.1 60.6 59.2
60.0
Payback [Years]
49.3
50.0 45.1
40.0
31.1
30.0
23.5 22.3
17.1 17.8
20.0
13.2 12.5 12.1 12.0
9.6 10.6 8.8 9.9 9.2 9.1
10.0 7.3
-
Hybrid ORC Hybrid, Stor Hybrid ORC Stor, Hybrid,
ORC ORC ORC
Technology option
Figure J2: Payback for 120 kW(e) Reference Plants
161
dni(i,1) = A(i,4); tstoras_clfr_steam(i,1)=thermstor(i,8);
end tstoras_clfr_orc(i,1)=thermstor(i,11);
APPENDIX K MODEL
% tdump_trough_steam_n(i,1)=thermstor(i,1);
start=1; % tdump_trough_orc_n(i,1)=thermstor(i,4);
INPUT finish=24; % tdump_trough_steam_h(i,1)=thermstor(i,3);
i=1; % tdump_trough_orc_h(i,1)=thermstor(i,6);
k=1; %
while i<row % tdump_clfr_steam_n(i,1)=thermstor(i,7);
DESIGN for i=start:finish % tdump_clfr_steam_h(i,1)=thermstor(i,9);
daynum(i)=k; % tdump_clfr_orc_n(i,1)=thermstor(i,10);
% tdump_clfr_orc_h(i,1)=thermstor(i,12);
format long; end %
clc; end
clear; start=finish+1;
finish=start+23; sumthermstor=zeros(365,col1);
k=k+1; % sumthermdump=zeros(365,col1);
T=csvread('C:\Users\user\Desktop\wits\design.csv'); end
j=1;
[row1,col1] = size(T); k=1; c=1;
thermout=zeros(8760,k); % for i = 1:(row)
i = 1; while c<(366)
for i = 1:(col1) %find thermal output from given dni j=c*24;
Asf(i) = T(2,i)*4; %Solar Field Area i=1;
nsf(i) = T(3,i); %solar field efficiency sumthermstor(c,2)=sum(tstoras_trough_steam((
parloss(i) = T(4,i); %parasitic losses col1 j-23):j));
npbn(i) = T(7,i); %net power block efficiency for i = 1:(row) sumthermstor(c,5)=sum(tstoras_trough_orc((j-
storloss(i)=T(6,i);%storage efficiency for k=1:col1 23):j));
piploss(i) = T(5,i); % PIPING/RECEIVER dnitherm(i,k)=dni(i)*Asf(1,k)/1000; sumthermstor(c,8)=sum(tstoras_clfr_steam((j-
EFFICIENCY 23):j));
% solcapf(i)=T(9,i);%solar capacity factor thermout(i,k)=dni(i)*Asf(1,k)*nsf(1,k)*parloss(1,k)* sumthermstor(c,11)=sum(tstoras_clfr_orc((j-
% actualcapf(i)=T(10,i); %total capacity factor piploss(1,k)*storloss(1,k)/1000; %kW 23):j));
end end %
end sumthermdump(c,1)=sum(tdump_trough_steam
_n((j-23):j));
%1 parabolic trough with steam cycle no storage, no %
hybridisation Eout=zeros(8760,k); sumthermdump(c,3)=sum(tdump_trough_steam
% 2parabolic trough with steam cycle+storage thermstor=zeros(8760,k); _h((j-23):j));
% 3parabolic trough with steam cycle+hybrid %
sumthermdump(c,4)=sum(tdump_trough_orc_n(
%4parabolic trough with orc no storage, no k=1; (j-23):j));
hybridisation i=1; %
%5parabolic trough with orc cycle +storage sumthermdump(c,6)=sum(tdump_trough_orc_h(
%6parabolic trough with orc cycle +hybrid row2=row*2; (j-23):j));
%
%7clfr trough with orc no storage, no hybridisation for i=1:(row) sumthermdump(c,7)=sum(tdump_clfr_steam_n(
%8clfr trough with orc cycle +storage for k=1:col1 (j-23):j));
%9clfr trough with orc cycle +hybrid %storage or dumping %
sumthermdump(c,9)=sum(tdump_clfr_steam_h(
%10clfr trough with orc no storage, no hybridisation if thermout(i,k)>thermmax(k) (j-23):j));
%11clfr trough with orc cycle +storage Eout(i,k)=thermmax(k)*npbn(1,k); %
%12clfr trough with orc cycle +hybrid thermstor(i,k)=(thermout(i,k)-thermmax(k)); sumthermdump(c,10)=sum(tdump_clfr_orc_n((j
-23):j));
%
%Common variables %min therm dumped sumthermdump(c,12)=sum(tdump_clfr_orc_h((j
Enet=480; %kW else if thermout(i,k)<thermmin(k) -23):j));
Eout(i,k)=0.0;
k=1; thermstor(i,k)=thermout(i,k); c=c+1;
for k=1:(col1) % end
thermmax(k)=Enet/(npbn(k)); else if thermmin(k)<thermout(i,k)<thermmax(k) end
thermmin(k)=thermmax(k)*0.25; %kW Eout(i,k)=thermout(i,k)*npbn(1,k);
end
end k=1;
end i=1;
A=csvread('C:\Users\user\Desktop\wits\joburg\date.
csv'); end %sum the storage and see how much electricity
end the thermal storage over the
%period of a day can produce. doesnt take into
[row, col] = size(A); end account max and min therm
%energy delivered to the powerblock. assumes
i = 1; %thermalstorage all storage heat is used to
for i = 1:(row) %produce electricity.
month(i,1) = A(i,1); i=1;
day(i,1) = A(i,2); for i=1:row for i=1:(c-1)
hour(i,1) = A(i,3); tstoras_trough_steam(i,1)=thermstor(i,2); for k=1:col1
tstoras_trough_orc(i,1)=thermstor(i,5);
162
Estor(i,k)=sumthermstor(i,k)*npbn(1,k); for k=1:col1 % end
Estor(i,col1)=i;
end dnithermc(odd,k)=dnitherm(i,k); %populating the date matrix
end dnithermc(even,k)=dnitherm(i,k); %year of interest,min(i)
for i = 1:row
thermoutc(odd,k)=thermout(i,k);
Estora=zeros(8760,12); thermoutc(even,k)=thermout(i,k); %l=[2008,monthcs(i),daycs(i),hourc(i),min(i),0]
;
% thermoutdump(odd,k)=thermdump l = [num2str(month(i)) '/' num2str(day(i)) '/'
for k=1:col1 num2str(2008) ' ' num2str(hour(i)) ':'
for i=1:row E(odd,k)=Eout(i,k); num2str(min(i)) ':' '00'];
for j=1:365 E(even,k)=Eout(i,k); DateNumber(i) = datenum(l);
if (daynum(i)==Estor(j,13)) && (hour(i) == 18) end
Estorac(odd,k)=Estora(i,k);
if Estor(j,k)<481 Estorac(even,k)=Estora(i,k);
Estora(i,k)=Estor(j,k); %%%%monthlythermal energy flow
Ehybridc(odd,k)=Ehybrid(i,k); p=1;
Ehybridc(even,k)=Ehybrid(i,k); n=1;
else if Estor(j,k)>481 && Estor(j,k)<960 for n=1:3
Estora(i,k)=480; end if n==1
Estora(i+1,k)=Estor(j,k)-480; dayc(odd,1)= day(i); for p=1:row
monthc(odd,1)=month(i); P1(p,1)=dnitherm(p,10);
else if Estor(j,k)>960&& Estor(j,k)<1440 hourc(odd,1)=hour(i); P1(p,2)=thermout(p,10);
Estora(i,k)=480; daynumc(odd,1)=daynum(i); P1(p,3)=Eout(p,10);
Estora(i+1,k)=480; end
dayc(even,1)= day(i); end
Estora(i+2,k)=Estor(j,k)-480; monthc(even,1)=month(i);
hourc(even,1)=hour(i); if n==2
else if Estor(j,k)>960 daynumc(even,1)=daynum(i); for p=1:row
Estora(i,k)=480; P2(p,4)=Estora(p,11);
Estora(i+1,k)=480; P2(p,1)=dnitherm(p,11);
Estora(i+2,k)=480; even = even+2; P2(p,2)=thermout(p,11);
Estora(i+3,k)=Estor(j,k)-1440; odd=odd+2; P2(p,3)=Eout(p,11);
end
end end
end end if n==3
end for p=1:row
end %rearrange to start with July1st P3(p,3)=Ehybrid(p,12);
Es=circshift(E,[8832,0]); %july 1 is on the 8689 th P3(p,1)=dnitherm(p,12);
end day: 17520-8688=8832 P3(p,2)=thermout(p,12);
%Estoras=circshift(Estora,[8832,0]); P3(p,4)=Eout(p,12);
end daycs=circshift(dayc,[8832,0]); end
end monthcs=circshift(monthc,[8832,0]); end
end thermoutcs=circshift(thermoutc,[8832,0]);
end daynumcs=circshift(daynumc, [8832,0]); end
Ehybridcs=circshift(Ehybridc,[8832,0]);
Estoras=circshift(Estorac,[8832,0]); %% Create figure
Ehybrid=zeros(row,col1); figure1 = figure;
Ewh=Es+Ehybridcs;
for i=1:row %% Create axes
for j=7:19 % for i=1:row2 axes1 = axes(...
% for k=1:col1 % Enot(i)=Es(i,1); 'XGrid','on',...
%while j>6 && j<21 % end 'XTick',[7.33408e+005 7.33468e+005
if hour(i)== j 7.33562e+005 7.33652e+005 7.33743e+005],...
Ehybrid(i,3)=480-Eout(i,3); %check
Ehybrid(i,6)=480-Eout(i,6); 'XTickLabel',{'01/01','01/03','01/06','01/09','01
Ehybrid(i,9)=480-Eout(i,9); %totals and capacity factors /12'},...
Ehybrid(i,12)=480-Eout(i,12); 'YGrid','on',...
sumEsolar=sum(Es)/2; 'Parent',figure1);
% end sumEstorout=sum(Estor); xlim(axes1,[7.33408e+005 7.33773e+005]);
end solartotal=sumEsolar+sumEstorout; title(axes1,'CLFR Hourly Energy Flow');
end xlabel(axes1,'Date');
end total=sum(Ewh)/2+sumEstorout; ylabel(axes1,'Energy [kW]');
box(axes1,'on');
solarcapacity=(solartotal/Enet)/(365*24); hold(axes1,'all');
E=zeros(10000,col1); totalcapacity=(total/Enet)/(365*24);
thermoutc=zeros(17520,col1); %% Create bar
Estorac=zeros(17520,12); %min=zeros(row2,1); bar1 = bar(...
i=1; DateNumber,P2(1:8760,1),...
i=1; while i<17520 'Parent',axes1,...
odd=1; min(i)=0.0; 'DisplayName','DNI Thermal Energy Received
even=2; min(i+1)=30.0; [kWth]',...
i=i+2; 'FaceColor',[0 0 0]);
while odd<(row2) end
% for i=1:row2 %% Create bar
for i=1:(row) % year(1:8832,1)=2007; bar2 = bar(...
% year(8833:17520,1)=2008; DateNumber,P2(1:8760,2),...
163
'Parent',axes1,... 'Parent',axes1,...
'DisplayName','Thermal energy Delivered to PB 'DisplayName','Energy [kW]',...
[kWth]',... 'EdgeColor',[1 1 0],... %%%%%%%%%%%%%%%%%%%%%%%
'EdgeColor',[1 0 0],... 'FaceColor',[1 1 0]); h=1;
'FaceColor',[1 0 0]); m=1;
while h<25
%% Create bar %% Create legend
bar3 = bar(... legend1 = legend(axes1,{'DNI Thermal Energy for i = 1:row
DateNumber,P2(1:8760,4),... Received [kWth]','Thermal energy Delivered to PB for k=1:col1
'Parent',axes1,... [kWth]','Enet Hybrid [kW]','Enet
'DisplayName','Enet Storage [kW]',... [kW]'},'Position',[0.5911 0.7628 0.3084 0.1557]); if hour(i) == h
'EdgeColor',[0 0 1],... dnithermhour(h,k) = dnithermhour(h,k) +
'FaceColor',[0 0 1]); %% Create figure dnitherm(i,k);
figure1 = figure; dnithermhouraverage=(dnithermhour)/(365);
%% Create bar
bar4 = bar(... %% Create axes Ehybridhour(h,k) = Ehybridhour(h,k) +
DateNumber,P2(1:8760,3),... axes1 = axes(... Ehybrid(i,k);
'Parent',axes1,... 'XGrid','on',... Ehybridhouraverage=(Ehybridhour)/(365);
'DisplayName','Enet [kW]',... 'XTick',[7.33408e+005 7.33468e+005
'EdgeColor',[1 1 0],... 7.33562e+005 7.33652e+005 7.33743e+005],... thermouthour(h,k) = thermouthour(h,k) +
'FaceColor',[1 1 0]); thermout(i,k);
'XTickLabel',{'01/01','01/03','01/06','01/09','01/12 thermouthouraverage=(thermouthour)/(365);
'},...
%% Create legend 'YGrid','on',... Ehour(h,k) = Ehour(h,k) + Eout(i,k);
legend1 = legend(axes1,{'DNI Thermal Energy 'Parent',figure1); Ehouraverage=(Ehour)/(365);
Received [kWth]','Thermal energy Delivered to PB xlim(axes1,[7.33408e+005 7.33773e+005]);
[kWth]','Enet Storage [kW]','Enet title(axes1,'CLFR Hourly Energy Flow'); Estorhour(h,k) = Estorhour(h,k) + Estora(i,k);
[kW]'},'Position',[0.5965 0.7548 0.3076 0.1684]); xlabel(axes1,'Date'); Estorhouraverage=(Estorhour)/(365);
ylabel(axes1,'Energy [kW]');
%% Create figure box(axes1,'on'); end
figure1 = figure; hold(axes1,'all'); end
%% Create axes %% Create bar end

axes1 = axes(... bar1 = bar(... h=h+1;
'XGrid','on',... DateNumber,P1(1:8760,1),... end
'XTick',[7.33408e+005 7.33468e+005 7.33562e+005 'Parent',axes1,...
7.33652e+005 7.33743e+005],... 'DisplayName','DNI Thermal Energy Received for m=1:24
[kWth]',... for k=1:12
'XTickLabel',{'01/01','01/03','01/06','01/09','01/12'},.. 'FaceColor',[0 0 0]); %
.
'YGrid','on',... %% Create bar Ehourtotal(m,k)=Ehouraverage(m,k)+Ehybridh
'Parent',figure1); bar2 = bar(... ouraverage(m,k);
xlim(axes1,[7.33408e+005 7.33773e+005]); DateNumber,P1(1:8760,2),...
title(axes1,'CLFR Hourly Energy Flow'); 'Parent',axes1,... Ehourtotal1(m,k)=Ehouraverage(m,k)+Estorhou
xlabel(axes1,'Date'); 'DisplayName','Thermal energy Delivered to PB raverage(m,k);
ylabel(axes1,'Energy [kW]'); [kWth]',... %
box(axes1,'on'); 'EdgeColor',[1 0 0],... end
hold(axes1,'all'); 'FaceColor',[1 0 0]); end
%% Create bar %% Create bar %thermal energy flow

bar1 = bar(... bar3 = bar(... n=1;
DateNumber,P3(1:8760,1),... DateNumber,P1(1:8760,3),... for n=1:3
'Parent',axes1,... 'Parent',axes1,... if n==1
'DisplayName','DNI Thermal Energy Received 'DisplayName','Enet [kW]',... for h=1:24
[kWth]',... 'EdgeColor',[1 1 0],... H1(h,3)=dnithermhouraverage(h,10);
'FaceColor',[0 0 0]); 'FaceColor',[1 1 0]); H1(h,2)=thermouthouraverage(h,10);
H1(h,1)=Ehourtotal(h,10);
%% Create bar end
bar2 = bar(... %% Create legend end
DateNumber,P3(1:8760,2),... legend1 = legend(axes1,{'DNI Thermal Energy
'Parent',axes1,... Received [kWth]','Thermal energy Delivered to PB if n==2
'DisplayName','Thermal energy Delivered to PB [kWth]','Enet [kW]'},'Position',[0.5911 0.7628 for h=1:24
[kWth]',... 0.3084 0.1557]);
'EdgeColor',[1 0 0],... H2(h,3)=dnithermhouraverage(h,11);
'FaceColor',[1 0 0]); H2(h,2)=thermouthouraverage(h,11);
H2(h,1)=Ehourtotal1(h,11);
%% Create bar end
bar3 = bar(... end
DateNumber,P3(1:8760,3),... %%%%%%%find the average data on an hourly if n==3
'Parent',axes1,... basis- this is for modelling the for h=1:24
'DisplayName','Enet Hybrid [kW]',... %%%%%%%energy conversion from collected dni H3(h,2)=Ehybridhouraverage(h,12);
'EdgeColor',[0 0 1],... to electricity H3(h,5)=dnithermhouraverage(h,12);
'FaceColor',[0 0 1]); dnithermhour=zeros(24,13); H3(h,4)=thermouthouraverage(h,12);
thermouthour=zeros(24,13); H3(h,1)=Ehouraverage(h,12);
%% Create bar Estorhour=zeros(24,13); H3(h,3)=Ehourtotal(h,12);
bar4 = bar(... Ehybridhour=zeros(24,13); end
DateNumber,P3(1:8760,4),... Ehour=zeros(24,13); end
164
end i=1; %this will be the year for which the bill is to be
B=zeros(row2,col1+3); determined
H=zeros(row2,col1+3); N=
%% Create figure csvread('C:\Users\user\Desktop\wits\Christiaa
figure1 = figure; for k=1:col1 n files\West
for i=1:(row2) Campus\West_July06_to_June07.csv');
%% Create axes %this is the previous year required to
axes1 = axes('YGrid','on','Parent',figure1); B(i,1)=monthcs(i); determine the bill
xlim(axes1,[0 23]);
title(axes1,'CLFR'); B(i,2)=daycs(i); E=
xlabel(axes1,'Time'); csvread('C:\Users\user\Desktop\wits\joburg\d
ylabel(axes1,'Energy [kW]'); B(i,3)=hourc(i); esignsolar.csv');
box(axes1,'on'); %S=
hold(axes1,'all'); B(i,4)=daynumcs(i); csvread('C:\Users\user\Desktop\wits\joburg\d
esignstorage.csv');
%% Create mutliple lines using matrix input to plot B(i,k+4)=Es(i,k); % H=
plot1 = plot(H1); csvread('C:\Users\user\Desktop\wits\joburg\d
set(plot1(2),'Color',[1 0 1]); end esignhybrid.csv');
end
%% Create legend L=csvread('C:\Users\user\Desktop\wits\joburg
legend1 = legend(axes1,{'Enet Electricity success = \lec.csv');
Generated','Thermal Energy Delivered','DNI Thermal xlswrite('C:\Users\user\Desktop\wits\joburgchoic
Energy'},'Position',[0.6797 0.7441 0.2583 0.1684]); e\designsolar.xls', B); [row,col] = size(M);
success =
xlswrite('C:\Users\user\Desktop\wits\joburgchoic [row_previous, col_previous] = size(N);
e\designstorage.xls', Estoras); [rows,cols] = size(E);
%% Create figure success =
figure1 = figure; xlswrite('C:\Users\user\Desktop\wits\joburgchoic
e\designhybrid.xls', Ehybridcs); LEC= L(1,1);
%% Create axes LECf=L(2,1);
axes1 = axes('YGrid','on','Parent',figure1); success = invest=L(3,1);
xlim(axes1,[0 23]); xlswrite('C:\Users\user\Desktop\wits\joburgchoic investf=L(4,1);
title(axes1,'CLFR with Storage'); e\designsolar.xls', B);
xlabel(axes1,'Time'); success = i = 1;
ylabel(axes1,'Energy [kW]'); xlswrite('C:\Users\user\Desktop\wits\joburgchoic for i = 1:(rows)
box(axes1,'on'); e\designstorage.xls', Estoras);
hold(axes1,'all'); success = solar(i)=E(i,5);
xlswrite('C:\Users\user\Desktop\wits\joburgchoic %Estoras(i)= S(i,5);
%% Create mutliple lines using matrix input to plot e\designhybrid.xls', Ehybridcs); end
plot1 = plot(H2);
set(plot1(2),'Color',[1 0 1]); %populating the year of interest's matrices
i = 1;
%% Create legend for i = 1:(row)
legend1 = legend(axes1,{'Enet Electricity BILL ANALYSIS Power(i,1) = M(i,1);
Generated','Thermal Energy Delivered','DNI Thermal Reactive(i,1) = M(i,2);
Energy'},'Position',[0.6797 0.7441 0.2583 0.1684]); Complex(i,1) = M(i,3);
Day(i,1) = M(i,4);
format long;
Month(i,1) = M(i,5);
clc;
% %%%%%%%%%%%%%%%% Year(i,1) = M(i,6);
clear;
Hour(i,1) = M(i,7);
%% Create figure Minute(i,1) = M(i,8);
%this code is written to read in two years of data.
figure1 = figure; end
Each year spans from
%July to June of the following year. This is in
%% Create axes %populating the previous year's matrices
conjunction with the
axes1 = axes('YGrid','on','Parent',figure1); i = 1;
%billing period of city power.
xlim(axes1,[0 23]); for i = 1:(row_previous)
%determination of the monthly and yearly cost
title(axes1,'CLFR with Hybridisation'); Power_previous(i,1) = N(i,1);
xlabel(axes1,'Time'); Reactive_previous(i,1) = N(i,2);
%service charge
ylabel(axes1,'Energy [kW]'); Complex_previous(i,1) = N(i,3);
service_charge = 1194.14;
box(axes1,'on'); Day_previous(i,1) = N(i,4);
hold(axes1,'all'); Month_previous(i,1) = N(i,5);
%energy charge (current) (kWh)
Year_previous(i,1) = N(i,6);
energy_cost_summer = 0.2345;
%% Create mutliple lines using matrix input to plot Hour_previous(i,1) = N(i,7);
energy_cost_winter = 0.3468;
plot1 = plot(H3); Minute_previous(i,1) = N(i,8);
set(plot1(2),'Color',[1 0 1]); end
%demand charge (current) (kVA)
demand_cost_summer = 78.24;
%% Create legend %with solar reduction- assuming solar takes
demand_cost_winter = 80.67;
legend1 = legend(axes1,{'Enet Electricity away from kVa as well as kW
Generated','Hybrid Electricity Generated','Total for i=1:row
reactive_energy_cost = 0.0613;
Electricity','Thermal Energy Delivered','DNI Thermal Powers(i)=Power(i)-solar(i);
surcharge = 0.02;
Energy'},'Position',[0.6797 0.7441 0.2583 0.1684]); Complexs(i)=Complex(i)-solar(i);
%% % end
M=
%end
csvread('C:\Users\user\Desktop\wits\Christiaan
for i=1:row_previous
files\West Campus\West_July07_to_June08.csv');
k=1; Power_previouss(i)=Power(i)-solar(i);
165
Complex_previouss(i)=Complex_previous(i)-solar(i); sorted_combined_complex_power_matrix = end
end sort(combined_complex_power_matrix,1,'descend'); k = 1;
%this is the combined_complex_power_matrix end
sorted in descending order for each month k = 1;
[row, col] = size(Minute) beginning = 0; for j = 7:12
[row_previous, col_previous] = size(Minute_previous) ending = 0; for i = 1:row
if Month(i) == j
%populating the date matrix %temp matrix is all the complex power readings of combined_complex_power_matrixs(k,j+6)
%year of interest the previous twelve = Complexs(i);
for i = 1:row %months (w.r.t each month of interest) combined k = k + 1;
l = [num2str(Month(i)) '/' num2str(Day(i)) '/' into one large matrix which end
num2str(Year(i)) ' ' num2str(Hour(i)) ':' %is then sorted in descending order to obtain the end
num2str(Minute(i)) ':' '00']; three largest peaks of k = 1;
DateNumber(i,1) = datenum(l); %these twelve months of readings. end
end k = 1;
%previous year temp_matrix = zeros(24000,1); for j = 1:6
for i = 1:row_previous three_largest_peaks = zeros(3,12); for i = 1:row
l = [num2str(Month_previous(i)) '/' for i = 13:24 if Month(i) == j
num2str(Day_previous(i)) '/' beginning = 1; combined_complex_power_matrixs(k,j+18)
num2str(Year_previous(i)) ' ' ending = 2000; = Complexs(i);
num2str(Hour_previous(i)) ':' for j = i-12:i-1 k = k+1;
num2str(Minute_previous(i)) ':' '00']; temp_matrix(beginning:ending,1) = end
DateNumber_previous(i,1) = datenum(l); combined_complex_power_matrix(1:2000,j); end
end beginning = ending + 1; k = 1;
ending = beginning + 1999; end
%combined_complex_power_month_matrix is used to end
determine the three largest temp_matrix = sort(temp_matrix, 'descend'); sorted_combined_complex_power_matrixs =
%peaks of the preceding twelve months for each three_largest_peaks(1:3,i-12) = sort(combined_complex_power_matrixs,1,'desce
month of interest. The fist temp_matrix(1:3,1); nd'); %this is the
%twelve columns is the end combined_complex_power_matrix sorted in
combined_complex_power_matrix = zeros(2000, 24); descending order for each month
k = 1; three_largest_peaks_average = zeros(1,12) beginning = 0;
for j = 7:12 %80% average of the three highest peaks w.r.t. each ending = 0;
for i = 1:row_previous of the months of
if Month_previous(i) == j %interest
combined_complex_power_matrix(k,j-6) = for i = 1:12 %temp matrix is all the complex power readings
Complex_previous(i); three_largest_peaks_average(i) = of the previous twelve
k = k+1; 0.8*(three_largest_peaks(1,i) + %months (w.r.t each month of interest)
end three_largest_peaks(2,i) + combined into one large matrix which
end three_largest_peaks(3,i))/3; %is then sorted in descending order to obtain
k = 1; end the three largest peaks of
end %these twelve months of readings.
k = 1; %extraction of the peak complex power of each temp_matrix = zeros(24000,1);
for j = 1:6 month of interest. The three_largest_peakss = zeros(3,12);
for i = 1:row_previous %twelve months of interest correspond to columns for i = 13:24
if Month_previous(i) == j 13 to 24 of the beginning = 1;
combined_complex_power_matrix(k,j+6) = %combined_complex_power_matrix. ending = 2000;
Complex_previous(i); peaks = sorted_combined_complex_power_matrix(1, for j = i-12:i-1
k = k+1; 13:24); temp_matrix(beginning:ending,1) =
end combined_complex_power_matrixs(1:2000,j);
end %%%%%solar peaks beginning = ending + 1;
k = 1; %combined_complex_power_month_matrix is used ending = beginning + 1999;
end to determine the three largest end
k = 1; %peaks of the preceding twelve months for each temp_matrix = sort(temp_matrix, 'descend');
for j = 7:12 month of interest. The fist three_largest_peakss(1:3,i-12) =
for i = 1:row %twelve columns is the temp_matrix(1:3,1);
if Month(i) == j combined_complex_power_matrixs = zeros(2000, end
combined_complex_power_matrix(k,j+6) = 24);
Complex(i); k = 1;
k = k + 1; for j = 7:12 three_largest_peaks_averages = zeros(1,12)
end for i = 1:row_previous %80% average of the three highest peaks w.r.t.
end if Month_previous(i) == j each of the months of
k = 1; combined_complex_power_matrixs(k,j-6) = %interest
end Complex_previouss(i);
k = 1; k = k+1; for i = 1:12
for j = 1:6 end three_largest_peaks_averages(i) =
for i = 1:row end 0.8*(three_largest_peakss(1,i) +
if Month(i) == j k = 1; three_largest_peakss(2,i) +
combined_complex_power_matrix(k,j+18) = end three_largest_peakss(3,i))/3;
Complex(i); k = 1; end
k = k+1; for j = 1:6
end for i = 1:row_previous
end if Month_previous(i) == j %extraction of the peak complex power of each
k = 1; combined_complex_power_matrixs(k,j+6) = month of interest. The
end Complex_previouss(i); %twelve months of interest correspond to
k = k+1; columns 13 to 24 of the
end %combined_complex_power_matrix.
166
peakss = sorted_combined_complex_power_matrixs(1, energy_monthss(2) = energy_monthss(2) + j = j+1;
13:24); Powers(i)*0.5; end
elseif Month(i) == 9 end
%determine the energy consumption for each month energy_monthss(3) = energy_monthss(3) + end
of the year of interest Powers(i)*0.5;
%this is correct. Checked against known data elseif Month(i) == 10 %determining the demand chargable for each
energy_months = zeros(1,12); energy_monthss(4) = energy_monthss(4) + month is calculated by using the
for i = 1:row Powers(i)*0.5; %greater of: 1) the measured demand, 2) a
if Month(i) == 1 elseif Month(i) == 11 demand of 70kVA 3) a demand based
energy_months(7) = energy_months(7) + energy_monthss(5) = energy_monthss(5) + %on the 80% average of the three highest
Power(i)*0.5; Powers(i)*0.5; demands recorded over the
elseif Month(i) == 2 elseif Month(i) == 12 %preceding 12 months.
energy_months(8) = energy_months(8) + energy_monthss(6) = energy_monthss(6) + demand = zeros(1,12);
Power(i)*0.5; Powers(i)*0.5; for i=1:12
elseif Month(i) == 3 end demand(i) = peaks(i);
energy_months(9) = energy_months(9) + end if three_largest_peaks_average(i) >
Power(i)*0.5; demand(i)
elseif Month(i) == 4 %determination of the kVArh consumption for each demand(i) =
energy_months(10) = energy_months(10) + month of the year of interest three_largest_peaks_average(i);
Power(i)*0.5; %correct. checked against known data elseif 70 > demand(i)
elseif Month(i) == 5 reactive_energy_months = zeros(1,12); demand(i) = 70;
energy_months(11) = energy_months(11) + for i = 1:row end
Power(i)*0.5; if Month(i) == 1 end
elseif Month(i) == 6 reactive_energy_months(7) =
energy_months(12) = energy_months(12) + reactive_energy_months(7) + Reactive(i)*0.5; %%%%%%%% solar
Power(i)*0.5; elseif Month(i) == 2
elseif Month(i) == 7 reactive_energy_months(8) = %determining the demand chargable for each
energy_months(1) = energy_months(1) + reactive_energy_months(8) + Reactive(i)*0.5; month is calculated by using the
Power(i)*0.5; elseif Month(i) == 3 %greater of: 1) the measured demand, 2) a
elseif Month(i) == 8 reactive_energy_months(9) = demand of 70kVA 3) a demand based
energy_months(2) = energy_months(2) + reactive_energy_months(9) + Reactive(i)*0.5; %on the 80% average of the three highest
Power(i)*0.5; elseif Month(i) == 4 demands recorded over the
elseif Month(i) == 9 reactive_energy_months(10) = %preceding 12 months.
energy_months(3) = energy_months(3) + reactive_energy_months(10) + Reactive(i)*0.5; demands = zeros(1,12);
Power(i)*0.5; elseif Month(i) == 5 for i=1:12
elseif Month(i) == 10 reactive_energy_months(11) = demands(i) = peakss(i);
energy_months(4) = energy_months(4) + reactive_energy_months(11) + Reactive(i)*0.5; if three_largest_peaks_averages(i) >
Power(i)*0.5; elseif Month(i) == 6 demands(i)
elseif Month(i) == 11 reactive_energy_months(12) = demands(i) =
energy_months(5) = energy_months(5) + reactive_energy_months(12) + Reactive(i)*0.5; three_largest_peaks_averages(i);
Power(i)*0.5; elseif Month(i) == 7 elseif 70 > demands(i)
elseif Month(i) == 12 reactive_energy_months(1) = demands(i) = 70;
energy_months(6) = energy_months(6) + reactive_energy_months(1) + Reactive(i)*0.5; end
Power(i)*0.5; elseif Month(i) == 8 end
end reactive_energy_months(2) =
end reactive_energy_months(2) + Reactive(i)*0.5; %%%%%%%%%%%%%%
elseif Month(i) == 9
%%%%%%%%solar months reactive_energy_months(3) = %determination of the excess reactive energy. A
reactive_energy_months(3) + Reactive(i)*0.5; charge will be made on the
%determine the energy consumption for each month elseif Month(i) == 10 %kVAh in excess of 30% of the kWh for each
of the year of interest reactive_energy_months(4) = month. Checked against known
%this is correct. Checked against known data reactive_energy_months(4) + Reactive(i)*0.5; %data
energy_monthss = zeros(1,12); elseif Month(i) == 11 billable_reactive_energy = zeros(1,12);
for i = 1:row reactive_energy_months(5) = for i = 1:12
if Month(i) == 1 reactive_energy_months(5) + Reactive(i)*0.5; if
energy_monthss(7) = energy_monthss(7) + elseif Month(i) == 12 (reactive_energy_months(i)/energy_months(i))
Powers(i)*0.5; reactive_energy_months(6) = > 0.3
elseif Month(i) == 2 reactive_energy_months(6) + Reactive(i)*0.5; billable_reactive_energy(i) =
energy_monthss(8) = energy_monthss(8) + end reactive_energy_months(i) -
Powers(i)*0.5; end 0.3*energy_months(i);
elseif Month(i) == 3 end
energy_monthss(9) = energy_monthss(9) + %determination of the months of interest end
Powers(i)*0.5; %extracting the months of interest into a matrix
elseif Month(i) == 4 from the Months matrix %determination of the montly and yearly
energy_monthss(10) = energy_monthss(10) + %extracting the year associated with each month of energy cost
Powers(i)*0.5; interest into a matrix energy_year = 0;
elseif Month(i) == 5 %from the years matrix energy_cost_year = 0;
energy_monthss(11) = energy_monthss(11) + Months_of_interest = zeros(1,12); energy_cost_months = zeros(1,12);
Powers(i)*0.5; Years_of_interest = zeros(1,12);
elseif Month(i) == 6 j = 1; for i = 1:12
energy_monthss(12) = energy_monthss(12) + for i = 1:row energy_year = energy_year +
Powers(i)*0.5; if Months_of_interest(j) == 0 energy_months(i);
elseif Month(i) == 7 Months_of_interest(j) = Month(i); if ((2 < i) && (i < 11)) %summer rates are
energy_monthss(1) = energy_monthss(1) + Years_of_interest(j) = Year(i); from September to April
Powers(i)*0.5; end energy_cost_months(i) =
elseif Month(i) == 8 if i < row energy_months(i)*energy_cost_summer;
if abs(Month(i) - Month(i+1)) > 0 value_i_1 = i
167
end if ((i < 3) || (i > 10)) %winter rates are from
if ((i < 3) || (i > 10)) %winter rates are from May to billable_reactive_energy_year = 0; %kVArh May to August
August billable_reactive_energy_cost_year = 0; demand_costs(i) = demands(i) *
energy_cost_months(i) = billable_reactive_energy_cost_months = zeros(1,12); demand_cost_winter;
energy_months(i)*energy_cost_winter; demand_2 = i
value_i_2 = i for i = 1:12 end
end billable_reactive_energy_year = end
energy_cost_year = energy_cost_year + billable_reactive_energy_year +
energy_cost_months(i); reactive_energy_months(i); %this is the determination of the max. demand
billable_reactive_energy_cost_months(i) = through the use of the peaks
end billable_reactive_energy(i)*reactive_energy_cost; %of each month. This should not be used
billable_reactive_energy_cost_year = according the the document on the
%%%%%%%solar year billable_reactive_energy_cost_year + %electricity tariff structure. However, it seems
billable_reactive_energy_cost_months(i); someone is billing wits
%determination of the montly and yearly energy cost %by taking the max. demand for each month as
energy_years = 0; end the peaks for each month.
energy_cost_years = 0; peaks_costs = zeros(1,12);
energy_cost_monthss = zeros(1,12);
%determination of the monthly and the yearly
demand costs for i = 1:12
for i = 1:12 demand_cost = zeros(1,12); if ((2 < i) && (i < 11)) %summer rates are
energy_years = energy_years + energy_monthss(i); from September to April
if ((2 < i) && (i < 11)) %summer rates are from demand_cost_year = 0; peaks_costs(i) = peakss(i) *
September to April demand_cost_summer;
energy_cost_monthss(i) = for i = 1:12 peaks_demand_1 = i
energy_monthss(i)*energy_cost_summer; if ((2 < i) && (i < 11)) %summer rates are from end
value_i_1 = i September to April if ((i < 3) || (i > 10)) %winter rates are from
end demand_cost(i) = demand(i) * May to August
if ((i < 3) || (i > 10)) %winter rates are from May to demand_cost_summer; peaks_costs(i) = peakss(i) *
August demand_1 = i demand_cost_winter;
energy_cost_monthss(i) = end peaks_demand_2 = i
energy_monthss(i)*energy_cost_winter; if ((i < 3) || (i > 10)) %winter rates are from May end
value_i_2 = i to August end
end demand_cost(i) = demand(i) *
energy_cost_years = energy_cost_years + demand_cost_winter;
energy_cost_monthss(i); demand_2 = i %%%%%%Find the total cost
end powercostsave=energy_cost_months-
end end energy_cost_monthss;
%%%%%%%%%%%%%%%% demandcostsave=demand_cost-demand_costs;
%this is the determination of the max. demand
through the use of the peaks for i=1:12
%determination of the montly and yearly energy cost %of each month. This should not be used according cfmonth(i)=solarmonth(i)/(120*24*30);
energy_years = 0; the the document on the actualLEC(i)=(solarcostmonth(i)-
energy_cost_years = 0; %electricity tariff structure. However, it seems (powercostsave(i)+demandcostsave(i)))/solarm
someone is billing wits onth(i);
energy_cost_monthss = zeros(1,12); %by taking the max. demand for each month as the actualLECf(i)=(solarcostmonthf(i)-
peaks for each month. (powercostsave(i)+demandcostsave(i)))/solarm
solarenergyyear=0; peaks_cost = zeros(1,12); onth(i);
solarcostyear=0; end
solarcostmonth=zeros(1,12);
for i = 1:12 averagecf=(sum(cfmonth))/12;
for i = 1:12 if ((2 < i) && (i < 11)) %summer rates are from averageLEC=(sum(actualLEC))/12;
energy_years = energy_years + energy_monthss(i); September to April averageLECf=(sum(actualLECf))/12;
if ((2 < i) && (i < 11)) %summer rates are from peaks_cost(i) = peaks(i) * demand_cost_summer;
September to April peaks_demand_1 = i %%%%%%%%%%%%%%
energy_cost_monthss(i) = end
energy_monthss(i)*energy_cost_summer; if ((i < 3) || (i > 10)) %winter rates are from May %computation of the total monthly bills. The
value_i_1 = i to August total cost per month will
end peaks_cost(i) = peaks(i) * demand_cost_winter; %consist of the energy cost, demand charge,
if ((i < 3) || (i > 10)) %winter rates are from May to peaks_demand_2 = i reactive energy charge and
August end %service charge
energy_cost_monthss(i) = end cost = zeros(1,12);
energy_monthss(i)*energy_cost_winter; surcharge_cost = zeros(1,12);
value_i_2 = i %%%%%%%%%%%%%%%solar demand tax = zeros(1,12);
end total_cost = zeros(1,12);
energy_cost_years = energy_cost_years + %determination of the monthly and the yearly
energy_cost_monthss(i); demand costs for i = 1:12
demand_costs = zeros(1,12); cost(i) = energy_cost_months(i) +
solarmonth(i)=energy_months(i)-energy_monthss(i); demand_cost_years = 0; demand_cost(i) +
solarcostmonth(i)=LEC*solarmonth(i); billable_reactive_energy_cost_months(i) +
solarcostmonthf(i)=(LEC-fit)*solarmonth(i); for i = 1:12 service_charge;
solarenergyyear= solarenergyyear+solarmonth(i); if ((2 < i) && (i < 11)) %summer rates are from surcharge_cost(i) = surcharge*cost(i);
solarcostyear=solarcostyear+solarcostmonth(i); September to April tax(i) = (cost(i) + surcharge_cost(i)) * 0.14;
demand_costs(i) = demands(i) * total_cost(i) = cost(i) + surcharge_cost(i) +
end demand_cost_summer; tax(i);
demand_1 = i
%%%%%%%%%% end end
168
summary(15,1)=sum(total_costssf)- 'XTickLabel',{'01-Jul-2007','01-Oct-2007','01-
%%%%%solar bill sum(total_cost);%extra cost of solar with fit Jan-2008','01-Apr-2008','01-Jul-2008'},...
summary(16,1)=paybackf; 'YGrid','on',...
%computation of the total monthly bills. The total cost 'YMinorTick','on',...
per month will success = 'Parent',figure1);
%consist of the energy cost, demand charge, reactive xlswrite('C:\Users\user\Desktop\wits\joburg\1\su axis(axes1,[7.33224e+005 7.33590e+005 -200
energy charge and mmary.xls', summary); 3000]);
%service charge success = title(axes1,'West Campus Power Usage
costs = zeros(1,12); xlswrite('C:\Users\user\Desktop\wits\joburg\1\re 2007/2008');
surcharge_costs = zeros(1,12); sult.xls', result); xlabel(axes1,'Date');
success = ylabel(axes1,'Power [kW]');
taxs = zeros(1,12); xlswrite('C:\Users\user\Desktop\wits\joburg\1\bil box(axes1,'on');
l.xls', bill); hold(axes1,'all');
total_costs = zeros(1,12);
%Hourly usage %% Create bar
total_costss= zeros(1,12); bar1 = bar(...
%% Create figure DateNumber,Power,...
for i = 1:12 figure1 = figure; 'Parent',axes1,...
costs(i) = energy_cost_monthss(i) + demand_costs(i) 'DisplayName','Power',...
+ billable_reactive_energy_cost_months(i) + %% Create axes 'BarLayout','stacked',...
service_charge; axes1 = axes(... 'FaceColor',[0 0 0]);
surcharge_costs(i) = surcharge*costs(i); 'XGrid','on',...
taxs(i) = (costs(i) + surcharge_costs(i)) * 0.14; 'XMinorTick','on',... %% Create bar
total_costs(i) = costs(i) + surcharge_costs(i) + taxs(i); 'XTick',[7.33224e+005 7.333e+005 7.334e+005 bar2 = bar(...
total_costss(i)=total_costs(i)+solarcostmonth(i); 7.335e+005 7.33590e+005],... DateNumber,Powers,...
'XTickLabel',{'01-Jul-2007','01-Oct-2007','01-Jan- 'Parent',axes1,...
total_costssf(i)=total_costs(i)+solarcostmonthf(i); 2008','01-Apr-2008','01-Jul-2008'},... 'DisplayName','Power with Solar',...
end 'YGrid','on',... 'BarLayout','stacked',...
'YMinorTick','on',... 'EdgeColor',[1 0 0],...
%%%%%%%%%%% 'Parent',figure1); 'FaceColor',[1 0 0]);
axis(axes1,[7.33224e+005 7.33590e+005 -200
result=zeros(18,13); 3000]); %% Create legend
payback=invest/(sum(total_cost)-sum(total_costs)); title(axes1,'West Campus Power Usage legend1 = legend(axes1,{'Power','Power with
paybackf=invest/(sum(total_cost)- 2007/2008'); Solar'});
(sum(total_costs))+fit*sum(solarmonth)); xlabel(axes1,'Date');
ylabel(axes1,'Power [kW]'); %%%%%total bill
result(2,2:13) = Years_of_interest; box(axes1,'on');
result(3,2:13) = Months_of_interest; hold(axes1,'all'); %% Create figure
result(4,2:13) = energy_months; figure1 = figure;
result(5,2:13) = peaks; %% Create bar
result(6,2:13) = demand; bar1 = bar(... %% Create axes
result(7,2:13) = reactive_energy_months; DateNumber,Power,... axes1 =
result(8,2:13) = billable_reactive_energy; 'Parent',axes1,... axes('XTickLabel',{'07/07','08/07','09/07','10/0
result(9,2:13)=solarmonth; 'DisplayName','Power',... 7','11/07','12/07','01/08','02/08','03/08','04/0
result(10,2:13)=cfmonth; 'BarLayout','stacked',... 8','05/08','06/08'},'Parent',figure1);
'FaceColor',[0 0 0]); xlim(axes1,[0.5 12.5]);
%bill title(axes1,'West Campus Total Bill ');
bill(2,2:13) = Years_of_interest; %% Create bar xlabel(axes1,'Date');
bill(3,2:13) = Months_of_interest; bar2 = bar(... ylabel(axes1,'Total Bill Cost [Rand]');
bill(4,2:13)=energy_cost_months; DateNumber,Powers,... box(axes1,'on');
bill(5,2:13)=energy_cost_months; 'Parent',axes1,... hold(axes1,'all');
bill(6,2:13)=demand_cost; 'DisplayName','Power with Solar',...
bill(7,2:13)=demand_costs; 'BarLayout','stacked',... %% Create bar
bill(8,2:13)=billable_reactive_energy_cost_months; 'EdgeColor',[1 0 0],... bar1 = bar(total_costss,...
bill(9,2:13)=total_cost; 'FaceColor',[1 0 0]); 'Parent',axes1,...
bill(10,2:13)=total_costs; 'BarLayout','stacked',...
bill(11,2:13)=total_costss; %% Create legend 'DisplayName','Solar Addition',...
bill(12,2:13)=actualLEC; legend1 = legend(axes1,{'Power','Power with 'FaceColor',[1 0 0]);
Solar'});
summary=zeros(17,1); %% Create bar
summary(2,1)=sum(energy_months); %%%%%%%%%%%%%day bar2 = bar(total_cost,...
summary(3,1)=sum(solarmonth); 'Parent',axes1,...
summary(4,1)=sum(total_cost); %% Create figure 'BarLayout','stacked',...
summary(5,1)=sum(total_costss); figure1 = figure; 'DisplayName','Normal Bill',...
summary(7,1)=sum(total_cost)-sum(total_costs); 'FaceColor',[0 0 1]);
summary(6,1)=sum(total_costss)-sum(total_cost); %% Create axes
summary(8,1)=LEC; axes1 = axes(... %% Create legend
summary(9,1)=averageLEC; 'XGrid','on',... legend1 = legend(axes1,{'Solar
summary(10,1)=averagecf; 'XMinorTick','on',... Addition','Normal Bill'},'Position',[0.4709 0.7746
summary(11,1)=invest; 'XTick',[7.33224e+005 7.333e+005 7.334e+005 0.1511 0.08917]);
summary(12,1)=payback; 7.335e+005 7.33590e+005],...
summary(14,1)=averageLECf;
169
APPENDIX L COMPARISON MODEL
The following table is the results from the Microsoft Excel model drawn up that compares
the Ecostar, Eskom and MTPP comparison under common assumptions. The
development of this model is given in Section 3.2 and the LECs are compared under
similar local conditions for the technology screening outlined in Section 4.1. The first four
pages of tables given here are the results from the Ecostar and MTPP analysis and the last
two from Eskom. As described, various scaling factors are extracted and applied to the
technologies allowing for scaling. Common DNI assumptions are also applied. Through
this, annual electricity generation can be verified and localised. Present day cost
assumptions are given by applying inflating costs according to the Chemical Engineering
Plant Cost Index.
170
COMPARISON MODEL PT with Thermal Oil PTwith DSG CLFR CRS Salt CRS Steam
given Scaled Plant given Scaled Plant given Scaled Plant Given sf Scaled Plant Given sf Scaled Plant
Design Electrical Output 50 100 47 100 50 100 17 51 100 11 55 100
Solar Field
Aperture Area of Solar Field [m2] 442035 884070.4094 448191 953598.3533 376200 752,400.35 152720 458160 1 898,352.94 93006 465032 1.00 845,514.08
Total Area of Plant[km2] 1.72 3.453128189 1.6 3.418409243 0.5643 1.13290711 0.611 1.833 1 3.594117647 0.372 1.86 1.00 3.38
0.256997093 0.256020154 0.280119375 0.278959681 0.666666667 0.664132424 0.2499509 0.2499509 0 0.2499509 0.250016129 0.250017204 1E-06 0.250017604
Area solar field for Adjusted 306,985.69 613,971.66 311,260.92 662,257.62 261,264.41 522,529.07 106,061.41 318,184.22 0.69 623,890.63 64,591.06 322,956.71 0.69 587,194.95
Total Area of Plant[km2] Adjusted 1,194,510.34 2,398,137.99 1,111,172.41 2,374,026.28 391,896.62 786,784.46 424,328.97 1,272,986.90 694,482.76 2,496,052.74 258,347.59 1,291,737.93 694,482.76 2,348,614.42
Lenght of Single Collector [m] 150 150 150 150 1000 1,000.00 121.34 121.34 0 121.34 121.34 121.34 0.00 121.34
Focal Length [m] 2.12 2.12 2.12 2.12 - 0 #DIV/0! #DIV/0!
Collector Row Spacing/Aperature Width 3 5.991031882 3 6.372587207 - 1259 3776 0.999758973 7402.720051 766 3776 0.99 6829.34
Average Reflectivity 0.88 0.88 0.88 0.88 0.88 0.88 0.88 0.88 0 0.88 0.88 0.88 0.00 0.88
Optical Peak Efficiency 0.75 0.75 0.75 0.75 0.64 0.64 0.75 0.75 0 0.75 0.75 0.75 0.00 0.75
HTF Temp at entrance [C] 291 291 126 126 126 126.00 0 0 0 0 0 #DIV/0! #DIV/0!
HTF Temp at exit [C] 391 391 411 411 411 411.00 0 0 0 0 0 #DIV/0! #DIV/0!
Factor for Solar field Parasitics [kW/m] 0.0098 0.0098 0 0.009 0.01 0 0 0 0.116 0.0016 -2.66 0.00
design parasitics for pumping and Tracking [kW] 4332 4332 4034 4034 3386 3,386.00 2482 7445 0.999877746 14596.83757 1490 7445 1.00 13532.99
Factor for Power Block Parasitics 0.03 0.03 0 0.03 0.03 0 0 0 0 0 #DIV/0! #DIV/0!
Operataing Mode 0 0 - 0 0 0 0 0 #DIV/0! #DIV/0!
Heat Loss factor piping [W/m2] 0.02 0.02 0.02 0.02 0.02 0.02 - - #VALUE! - - #VALUE! #VALUE!
Concentrator efficiency 54.20% 0.542 54.20% 0.542 42.20% 0.42 61.00% 61.00% 0 0.61 57.00% 57.00% 0.00 0.57
Efficiency loss due to parasitics 90.80% 0.908 89.90% 0.899 90.90% 0.91 85.00% 85.00% 0 0.85 96.00% 96.00% 0.00 0.96
Power Block Design

Design net Electrical Otput [kW] 50000 100000 47000 100000 50000 100,000.00 17000 51000 1 100000 11000 55000 1.00 100000.00
Design Efficiency of Power Block 0.375 0.375 26% 0.26 39% 0.39 38% 38% 0 0.38 30% 30% 0.00 0.30
Storage Capacity [h] 3 3 0 0 0 - 3 3 0 3 0.4 0.4 0.00 0.40
Thermal Capacity of the Storage [kWh] 434656 869312 0 0 0 - 153803 461409 1 904723.5294 14718 73590 1.00 133800.00
HTF temp in Storage Discharging [C] 371 371 0 0 0 - 560 560 0 560 260 260 0.00 260.00
Efficiency Factor dur to lower storage fluid Temp 0.975 0.975 0 0 0 - 0.997 0.997 0 0.997 0.8 0.8 0.00 0.80
overall Plant Availibility 0.96 0.96 0.96 0.96 0.96 0.96 0.96 0.96 0 0.96 0.96 0.96 0.00 0.96
Power Block Efficiency (incl. availibility and dumping) 35.30% 0.353 23% 0.228 32.80% 0.33 33% 33% 0 0.33 28% 28% 0.00 0.28
Storage Efficiency 94.70% 0.947 100% 1 100% 1.00 95% 95% 0 0.95 100% 100% 0.00 1.00
Receiver
design solar thermal input tp receiver - - - 73993 221979 1 435252.9412 45062 225308 1.00 409649.56
Max. Temp at receiver exit - - - 565 565 0 565 260 260 0.00 260.00
Receiver/Piping Efficiency 85.10% 0.851 89.20% 0.892 83.80% 0.84 84.00% 84% 0 0.84 88% 88% 0.00 0.88
O&M Input
Labour costs per employee 67,087.06 67087.06457 67,087.06 67087.06457 67,087.06 67,087.06 67,087.06 67,087.06 0 67087.06457 67,087.06 67,087.06 0.00 67087.06
Specific number of persons for field maintenance [/m2] 0.03 0.03 0.03 0.03 0.03 0.03 0.03 0.03 0 0.03 0.03 0.03 0.00 0.03
number of persons 30 30 30 30 30 30.00 30 30 0 30 30 30 0.00 30.00
number of persons for field maintenance 13 26.15096914 13 27.83456202 7.5 15.09 4.6 13.7 0.993380049 26.74327068 2.8 14 1.00 25.45
O&M Equipment cost percentage of investment[per a] 1% 0.01 1% 0.01 1% 0.01 1% 1% 0 0.01 1% 1% 0.00 0.01
power block O&M fix [*/kW] 27.00 27 27.00 27 27.00 27.00 27.00 27.00 0 27 27.00 27.00 0.00 27.00
powerblock O&M Variable[*/MWh] 2.50 2.5 2.50 2.5 2.50 2.50 2.50 2.50 0 2.5 2.60 2.60 0.00 2.60
Water Costs 0 0 - 0 1.3 1.3 0.00 1.30
Investment
Specific Investment cost for Solar Field (*/m2) 287.92 277.9502128 265.55 255.5570116 167.72 161.91 209.65 198.47 -0.049888607 191.9097236 209.65 192.88 -0.05 186.99
Specific Investment cost for Power Block [*/kWel] 978.35 931.8099781 607.98 576.5393333 978.35 931.81 1,048.24 969.97 -0.070635698 924.9133283 888.90 793.86 -0.07 761.21
Specific Land Cost [*/m2] 2.80 2.795294357 2.80 2.795294357 2.80 2.80 2.80 2.80 0 2.795294357 2.80 2.80 0.00 2.80
Specific Investment Storage[*/kWhth] 43.33 41.31959751 0.00 0 0.00 - 19.57 18.17 -0.067455983 17.36260012 139.76 124.39 -0.07 119.12
Total investment cost for tower 0.00 0 0.00 0 0.00 - 2,795,294.36 7,765,158.61 0.930000034 14524797.19 2,795,294.36 12,487,331.83 0.93 21773705.84
Specific investment cost for receiver [*/kWh] 0.00 0 0.00 0 0.00 - 174.71 162.13 -0.068016303 154.8693751 153.74 136.97 -0.07 131.22
Surcharge for Construction, engineering and Contingencies % 20% 0.2 20% 0.2 20% 0.20 20% 20% 0 0.2 20% 20% 0.00 0.20
PT with Thermal Oil PTwith DSG CLFR CRS Salt CRS Steam
Financial Parameters 0 0 - #DIV/0! #DIV/0! #DIV/0! #DIV/0!
Annual Insurance 1% 0.01 1% 0.01 1% 0.01 1% 1% 0 0.01 1% 1% 0.00 0.01
Lifetime 30 30 30 30 30 30.00 30 30 0 30 30 30 0.00 30.00
Debt Interest Rate 8% 0.08 8% 0.08 8% 0.08 8% 8% 0 0.08 8% 8% 0.00 0.08
Economic Results
Fixed Charge Rate 9.88% 9.88% 9.88% 9.88% 9.88% 9.88% 9.88% 9.88% 9.88% 9.88% 9.88% 9.88% 9.88% 9.88%
Investment Solar Field 88,385,881.70 170653552.6 119018448.5 243698745.4 63,095,384.23 121,823,205.99 32017301.57 90929136.45 0.950111393 172402664.7 19498436.02 89693191.46 0.95 158105222.66
Investment Power Block, BOP 54,623,839.36 104073489.9 31,884,551.64 64347148.53 53,698,177.63 102,309,848.87 20,956,186.22 58,215,012.72 0.92999996 108891685.7 10,203,653.21 45,582,465.35 0.93 79480486.27
Investment receiver 0 0 0 0 0 - 12,926,985.51 35,910,382.09 0.929999994 67170682.45 6,927,829.58 30,948,480.24 0.93 53963739.39
Investment Tower 0 0 0 0 0 - 2,795,294.36 7,765,158.61 0.930000034 14,524,797.19 2,795,294.36 12,487,331.83 0.93 21,773,705.84
Investment Storage 18,832,367.69 35919632.89 0 0 0 - 3009472.606 8383530.831 0.932544017 15708352.86 2057057.117 9153904.172 0.93 15938382.14
Investment Land 3,339,008.03 6,703,501.59 3,106,053.98 6,636,102.27 1,095,466.41 2,199,294.16 1,186,124.36 3,558,373.09 1 6,977,202.13 722,157.55 3,610,787.75 1.00 6,565,068.64
Contingencies 33,036,219.36 63356045.75 30801810.81 62606376.93 23577805.65 45,216,933.50 14578272.92 40952318.76 0.940164405 77127740.43 8440885.567 38295232.16 0.94 67158592.71
Sum Total Equipment Costs 165,181,096.78 316780228.8 154009054.1 313031884.7 117889028.3 226,084,667.51 72891364.62 204761593.8 0.940164405 385638702.1 42204427.84 191476160.8 0.94 335792963.57
Total Including indirect Costs 198,217,316.14 380136274.5 184810864.9 375638261.6 141466833.9 271,301,601.01 87469637.55 245713912.5 0.940164405 462766442.6 50645313.4 229771393 0.94 402951556.28
Specific Investment 3964.346323 3801.362745 3932.146061 3756.382616 2829.336679 2,713.02 5145.272797 4817.919854 -0.059835595 4627.664426 4604.1194 4177.66169 -0.06 4029.52
Actual O&M Costs 5,595,466.50 16,186,704.95 4,912,908.73 15,995,173.62 4,083,448.46 11,552,380.72 3,959,377.92 7,713,438.67 0.607017758 19,705,206.71 3,040,029.37 6,957,192.75 0.51 17,158,209.80
O&M % 0.028 0.02 0.027 0.019984997 0.029 0.02 0.045 0.031 -0.333146647 0.02508397 0.060 0.030 -0.43 0.02
Annual Financing & insurance Costs 19589308.61 37567892.35 18264383.44 37123365.28 13980804.11 26,812,040.90 8,644,399.78 24283275.32 0.940164405 45734019.78 5005146.337 22707717.03 0.94 39822668.09
Annual Fuel Costs 0.00 0 0.00 0 0.00 - 0.00 0.00 #DIV/0! 0 0.00 0.00 #DIV/0! 0.00
O&M Cost/ kWh 0.04 0.033110923 0.06 0.039499858 0.05 0.04 0.08 0.05 -0.393408136 0.0397111 0.12 0.05 -0.48 0.04
Actual Net Elec 124,670,469.84 249407794.7 89,299,577.41 190054593.1 80,034,878.80 160,112,676.63 49,655,503.38 149,036,226.66 1.000425893 292311710.5 25,454,157.82 127,065,290.07 1.00 230889164.66
Solar Net Electricity(Adjusted) 124,639,444.16 249,279,003.76 89,450,192.02 190,319,653.49 79,886,327.05 159,772,728.08 41,997,261.61 125,991,784.82 1 247,042,715.34 25,255,572.93 126,278,407.75 1.00 229597471.79
Fossil net Electricity(actual )Seville 0.00 0.00 0.00 0.00 0.00 0.00 7,658,241.78 23,044,441.83 0.00 45,268,995.15 0.00 0.00 -0.00 0.00
Total Calculated Enet 124,639,444.16 249,279,003.76 89,450,192.02 190349720.2 79,886,327.05 159,772,728.08 49,655,503.38 149,036,226.66 1.000425893 292311710.5 25,255,572.93 126,278,407.75 1.00 229597471.79
Calculated LEC Adjusted 0.202 0.216 0.259 0.279 0.226 0.240 0.300 0.254 -0.1519878 0.26489033 0.319 0.235 -0.19 0.248
#DIV/0! #DIV/0! #DIV/0!
Actual Capacity Factor 29.00% 29.00% 22.00% 22.00% 18.30% 18.30% 33.00% 33.00% 0 33.00% 26.00% 26.00% 0.00 26.00%
total capacity 28.46% 28.46% 21.73% 21.73% 18.24% 18.24% 33.34% 33.36% 0.01% 33.37% 26.21% 26.21% 0.01% 26.21%
solar capacity factor for joburg 28.46% 28.46% 21.73% 21.73% 18.24% 18.24% 28.20% 28.20% 0.01% 28.20% 26.21% 26.21% 0.01% 26.21%
Specific Investment Rand 51,904.39 49,770.48 51,482.80 49,181.57 37,043.94 35,520.98 67,366.03 63,080.06 -0.78 60,589.08 60,280.81 54,697.29 -0.79 52,757.64
LEC Rand 2.65 2.82 3.39 3.65 2.96 3.14 3.93 3.33 3.47 4.17 3.08 3.25
CRS Air CRS Hybrid Dish stirling MTPP
Given sf Scaled Plant Given sf Scaled Plant given Scaled Plant given Scaled Plant Average Scaling Factor
Design Electrical Output 10 50 100 14.683 58.732 100 50 100 1 100
Solar Field
Aperture Area of Solar Field [m2] 104580 522900 1.00 1,045,800.00 38000 152000 1.00 258,802.70 350000 700,000.32 27182.52234 2,718,260.60 1.00
Total Area of Plant[km2] 0.418 2.092 1.00 4.19 0.432 1.78 1.02 3.07 1.4 2.81 - 1.01
0.250191388 0.249952199 9.99406E-07 0.249849256 0.087962963 0.085393258 0.084426835 0.25 0.249049659
Area solar field for Adjusted 72,629.01 363,145.03 0.69 726,290.07 26,390.34 105,561.38 0.69 179,734.01 243,068.97 486,138.16 18,877.79 1,887,785.12 0.69
Total Area of Plant[km2] Adjusted 290,293.79 1,452,857.93 694,895.49 2,906,913.07 300,016.55 1,236,179.31 709,335.67 2,128,873.02 972,275.86 1,951,972.78 - - 698,299.17
Lenght of Single Collector [m] 121.34 121.34 0.00 121.34 121.34 121.34 0.00 121.34 120.4 120.40 - 0.00
Focal Length [m] 0.00 0.00 - - 0.00
Collector Row Spacing/Aperature Width 862 4309 1.00 8,617.14 313 1253 1.00 2,134.07 2907 5,805.31 - 1.00
Average Reflectivity 0.88 0.88 0.00 0.88 0.88 0.88 0.00 0.88 0.88 0.88 0.9 0.90 0.00
Optical Peak Efficiency 0.75 0.75 0.00 0.75 0.75 0.75 0.00 0.75 0.75 0.75 0.78 0.78 0.00
HTF Temp at entrance [C] 0 0 0.00 0 0 0 0 0 - - 0.00
HTF Temp at exit [C] 0 0 0.00 0 0 0 0 0 - - 0.00
Factor for Solar field Parasitics [kW/m] 0.0065 0.0065 0.00 0.01 0 0 0 0 0 - - 0.00
design parasitics for pumping and Tracking [kW] 680 3399 1.00 6,797.14 0 0 0 0 0 - - 0.00
Factor for Power Block Parasitics 0.03 0.03 0.00 0.03 0 0 0 0 0 - - 0.00
Operataing Mode 0 0 0 0 0 0 0 - - 0.00
Heat Loss factor piping [W/m2] - - - - - - 0.00
Concentrator efficiency 61.00% 61.00% 0.00 0.61 50.90% 50.90% 0.00 0.51 88.00% 0.88 44.00% 0.44 0.00
Efficiency loss due to parasitics 93.00% 93.00% 0.00 0.93 100.00% 100.00% 0.00 1.00 100.00% 1.00 100.00% 1.00 0.00
-
Power Block Design
Design net Electrical Otput [kW] 10000 50000 1.00 100,000.00 14683 58732 1.00 100,000.00 50000 100,000.00 1000 100,000.00 1.00
Design Efficiency of Power Block 34% 34% 0.00 0.34 45% 45% 0.00 0.45 21% 0.21 - 0.00
Storage Capacity [h] 3 3 0.00 3.00 0 0 0.00 0.00 0 - 12 12.00 0.00
Thermal Capacity of the Storage [kWh] 94233 471166 1.00 942,332.86 0 0 0.00 0.00 0 - - 1.00
HTF temp in Storage Discharging [C] 650 650 0.00 650.00 0 0 0.00 0.00 0 - - 0.00
Efficiency Factor dur to lower storage fluid Temp 0.985 0.985 0.00 0.99 0 0 0.00 0.00 0 - - 0.00
overall Plant Availibility 0.96 0.96 0.00 0.96 0.96 0.96 0.00 0.96 0.96 0.96 0.99 0.99 0.00
Power Block Efficiency (incl. availibility and dumping) 31% 31% 0.00 0.31 40% 40% 0.00 0.40 21.30% 0.21 22.50% 0.23 0.00
Storage Efficiency 100% 100% 0.00 1.00 100% 100% 0.00 1.00 100% 100% 97.00% 0.97 0.00
0.5631 0.56 -
-
Receiver
design solar thermal input tp receiver 50669 253345 1.00 506,690.00 18500 74000 1.00 125,996.05 233125 466,249.55 - 1.00
Max. Temp at receiver exit 680 680 0.00 680.00 800 800 0.00 800.00 800 800.00 - 0.00
Receiver/Piping Efficiency 77% 77% 0.00 0.77 93.80% 93.80% 0.00 0.94 89.20% 0.89 88.00% 0.88 0.00
O&M Input
Labour costs per employee 67,087.06 67,087.06 0.00 67,087.06 67,087.06 67,087.06 0.00 67,087.06 67,087.06 67,087.06 - 0.00
Specific number of persons for field maintenance [/m2] 0.03 0.03 0.00 0.03 0.03 0.03 0.00 0.03 0.06 0.06 - 0.00
number of persons 30 30 0.00 30.00 30 30 0.00 30.00 30 30.00 - 0.00
number of persons for field maintenance 3.1 15.7 1.01 31.57 1.1 4.6 1.03 7.97 21 42.24 - 1.01
O&M Equipment cost percentage of investment[per a] 1% 1% 0.00 0.01 1% 1% 0.00 0.01 2% 0.02 - 0.00
power block O&M fix [*/kW] 27.00 27.00 0.00 27.00 27.00 27.00 0.00 27.00 40.00 40.00 - 0.00
powerblock O&M Variable[*/MWh] 2.50 2.50 0.00 2.50 2.50 2.50 0.00 2.50 4.50 4.50 - 0.00
Water Costs 1.30 1.30 0.00 1.30 - - 1.00
Investment
Specific Investment cost for Solar Field (*/m2) 209.65 192.88 -0.05 186.07 209.65 195.67 -0.05 190.56 614.96 593.68 195.62 154.80 -0.05
Specific Investment cost for Power Block [*/kWel] 838.59 749.14 -0.07 713.62 978.35 887.51 -0.07 854.92 4,192.94 3,993.47 1,662.76 1,202.79 -0.0703
Specific Land Cost [*/m2] 2.80 2.80 0.00 2.80 2.80 2.80 0.00 2.80 2.80 2.80 0.00 - 0.00
Specific Investment Storage[*/kWhth] 83.86 75.47 -0.07 72.12 0.00 0.00 0.00 0.00 - 17.21 12.56 -0.07
Total investment cost for tower 2,795,294.36 12,487,331.83 0.93 23,791,813.60 2,795,294.36 10,147,132.36 0.93 16,645,228.63 0.00 - 0.00 - 0.93
Specific investment cost for receiver [*/kWh] 160.73 143.96 -0.07 137.28 209.65 190.08 -0.07 183.06 167.72 159.80 0.00 - -0.07
Surcharge for Construction, engineering and Contingencies % 20% 20% 0.00 0.20 20% 20% 0.00 0.20 20% 0.20 0.10 0.10 0.00
CRS Air CRS Hybrid Dish stirling 0.00 -
Financial Parameters - 0.00 -
Annual Insurance 1% 1% 0.00 0.01 1% 1% 0.00 0.01 1% 0.01 0.00 - 0.00
Lifetime 30 30 0.00 30.00 30 30 0.00 30.00 30 30.00 0.00 - 0.00
Debt Interest Rate 8% 8% 0.00 0.08 8% 8% 0.00 0.08 8% 0.08 0.00 - 0.00
Economic Results 0.00

Fixed Charge Rate 9.88% 9.88% 9.88% 9.88% 9.88% 9.88% 9.88% 9.88% 9.88% 9.88% 9.88% 9.88% 9.88%
Investment Solar Field 21924891.29 100854499.9 0.95 194,594,043.37 7966588.918 29741931.96 0.95 49,316,447.41 215237665.5 415,576,238.75 5,317,421.61 420,790,865.86 0.95
Investment Power Block, BOP 9,207,590.60 41,132,784.42 0.93 78,369,307.16 14,365,157.47 52,146,622.95 0.93 85,540,667.68 209,647,076.78 399,435,543.00 0.00 - 0.93
Investment receiver 8,144,000.66 36,381,442.30 0.93 69,316,689.27 3,878,470.92 14,079,147.14 0.93 23,095,256.99 39,099,179.82 74,494,729.70 0.00 - 0.93
Investment Tower 2,795,294.36 12,487,331.83 0.93 23,791,813.60 2,795,294.36 10,147,132.36 0.93 16,645,228.63 0.00 0.00 0.00 0.93
Investment Storage 7902269.194 35560286.85 0.93 67,965,570.96 0 0 0 - 0.00 - 0.93
Investment Land 811,456.60 4,061,165.58 1.00 8,125,677.70 838,634.57 3,455,485.05 1.02 5,950,826.74 2,717,797.23 5,456,338.50 0.00 0.00 1.00
Contingencies 10157100.54 46095502.18 0.94 88,422,982.16 5968829.247 21914063.89 0.94 36,104,167.38 93340343.86 179,005,806.71 0.00 - 0.94
Sum Total Equipment Costs 50785502.7 230477510.9 0.94 442,114,910.78 29844146.23 109570319.5 0.94 180,520,836.92 466701719.3 895,029,033.53 0.00 - 0.94
Total Including indirect Costs 60942603.24 276573013.1 0.94 530,537,892.94 35812975.48 131484383.3 0.94 216,625,004.31 560042063.2 1,074,034,840.23 7,544,051.77 570,784,354.38 0.94
Specific Investment 6094.260324 5531.460262 -0.06 5,305.38 2439.077537 2238.717962 -0.06 2,166.25 11200.84126 10,740.35 7,544.05 5,707.84 -0.06
Actual O&M Costs 3,263,226.63 8,142,225.65 0.57 22,591,004.63 3,352,459.42 7,253,001.98 0.56 9,224,178.97 16,004,790.48 45,733,822.92 113,125.41 8,559,089.32 0.56
O&M % 0.054 0.029 -0.37 0.023 0.094 0.055 -0.38 0.05 0.029 0.02 0.015 0.015 -0.38
Annual Financing & insurance Costs 6022801.062 27333001.03 0.94 52,431,698.27 3539304.449 12994264.14 0.94 21,408,493.18 55347519.69 106,144,106.63 745,559.27 56,409,152.76 0.94
Annual Fuel Costs 0.00 0.00 #DIV/0! 0.00 3,013,613.83 12,054,455.34 1.00 20,524,510.21 13,827,816.63 27,655,633.27 0.00 - 1.00
O&M Cost/ kWh 0.11 0.06 -0.43 0.04 0.05 0.03 -0.44 0.02 0.07 0.05 0.02 0.00 -0.44
Actual Net Elec 28,498,432.62 142,885,841.65 1.00 286,111,449.37 70,901,245.14 283,766,613.30 1.00 483,260,721.69 217,171,191.09 434,458,841.14 4,626,353.57 463,460,118.67 1.00
Solar Net Electricity(Joburg) 28,521,562.00 142,607,810.02 1.00 285,215,620.04 14,615,836.46 58,463,345.83 1.00 99,542,576.16 117,856,798.75 235,713,706.65 4,626,353.57 462,636,780.49 1.00
Fossil net Electricity(actual )Seville 0.00 0.00 0.00 0.00 56,285,408.69 225,303,267.47 0.00 383,718,145.53 99,314,392.34 198,745,134.49 0.00 0.00 0.00
Total Calculated Enet 28,521,562.00 142,607,810.02 1.00 285,215,620.04 70,901,245.14 283,766,613.30 1.00 483,260,721.69 217,171,191.09 434,458,841.14 4,626,353.57 462,636,780.49 1.00
Calculated LEC Joburg 0.326 0.249 -0.17 0.263 0.140 0.114 -0.15 0.106 0.392 0.413 0.186 0.140 -0.16
#DIV/0!
Actual Capacity Factor 33.00% 33.00% 0.00 33.00% 11.00% 11.00% 0.00 0.11 22.00% 22.00% 53.00% 53.00% 0.00
total capacity 32.56% 32.56% 0.01% 32.56% 55.12% 55.15% 0.01% 55.17% 49.58% 49.60% 52.81% 52.81% 0.01%
solar capacity factor for joburg 32.56% 32.56% 0.01% 32.56% 11.36% 11.36% 0.01% 11.36% 26.91% 26.91% 52.81% 52.81% 0.01%
Specific Investment Rand 79,790.93 72,422.30 -0.79 69,462.27 31,934.35 29,311.09 -0.81 28,362.28 146,650.37 140,621.23 98,772.76 74,731.65 -0.79
LEC Rand 4.26 3.26 3.44 1.83 1.49 1.39 5.14 5.41 2.43 1.84 -2.15
Paraboli DISH
ESKOM sf PT Only PT Hybrid PT with Salt Storage PT with DSG ISCCS CLFR Coal
Project pilot future Scaled pilot future Scaled pilot future Scaled pilot future Scaled pilot future Scaled pilot future Scaled
Plant Size [Mwe] 100 200 100 100 200 100 100 200 100 100 200 100 30 100 100 30 100 100
Solar Field [m2 x1000] 1 589 1086 589 589 1086 589 831 2184 831 536 994 536 170 942 566.6666667 196 667 653.3333333
Adjusted solar Field 1 589.00 1,086.00 589.00 589.00 1,086.00 589.00 831.00 2,184.00 831.00 536.00 994.00 536.00 170.00 942.00 566.67 196.00 667.00 653.33
0% 0 0 0 0 0
Thermal Storage[hrs] 1 0 0 0% 0 0 0 4 10 4 0 0 0 0 10 0 0 0 0
Annual Solar CF 1 25% 25% 25% 25% 25% 0.25 35% 50% 0.35 25% 25% 0.25 27% 52% 25% 25% 26% 25%
Annual Solar/Elec % 1 14% 15% 15% 14% 15% 0.138 14% 15% 0.136 15% 16% 0.151 16% 18% 0.516666667 13% 13% 0.416666667
Solar Fraction % 1 100% 100% 100% 75% 75% 0.75 100% 100% 1 100% 100% 1 100% 80% 3.333333333 0
0 0 0 0
Capital Cost [euroM] 0 0 0 0
Infrastructure 0.94 9.55 8.19 9.553792528 9.55 8.19 9.553792528 12.28 12.28 12.28344754 9.55 8.19 9.553792528 6.82 8.19 21.14743431
Solar Field 0.95 165.14 173.33 165.144128 165.14 173.33 165.144128 223.83 323.46 223.8317107 156.96 165.14 156.955163 55.96 152.86 175.4563262
Adjusted Solar Field Cost 0.95 165.14 173.33 165.144 165.14 173.33 165.14 223.83 323.46 223.83 156.96 165.14 156.96 55.96 152.86 175.46 - - -
Tower/Receiver 0.93 0.00 0.00 0 0.00 0.00 0 0.00 0.00 0 0.00 0.00 0 0.00 0.00 0
HTF System 0.93 15.01 17.74 15.01310254 34.12 38.22 34.1206876 16.38 20.47 16.37793005 1.36 2.73 1.364827504 0.00 1.36 0
Storage 0.93 0.00 0.00 0 0.00 0.00 0 43.67 68.24 43.67448013 0.00 0.00 0 0.00 28.66 0
Adjusted Storage 0.93 - - - - - - 43.67 68.24 43.67 - - - - 28.66 - - - -
Power Block 0.93 46.40 45.04 46.40413514 46.40 45.04 46.40413514 46.40 45.04 46.40413514 46.40 45.04 46.40413514 5.46 8.19 16.72688189 0.00
Balance of Plant 0.94 27.30 25.93 27.29655008 27.30 25.93 27.29655008 27.30 25.93 27.29655008 27.30 25.93 27.29655008 0.00 0.00 0 0.00
Services 0.94 25.93 16.38 25.93172258 28.66 17.74 28.66137758 36.85 30.03 36.85034261 24.57 15.01 24.56689507 6.82 12.28 21.14743431 0.00
Land 1.00 1.36 2.73 1.364827504 1.36 2.73 1.364827504 1.36 5.46 1.364827504 1.36 2.73 1.364827504 0.00 2.73 0 0.00
Contingencies 0.94 43.67 15.01 43.67448013 46.40 15.01 46.40413514 61.42 25.93 61.41723768 39.58 13.65 39.57999762 12.28 10.92 38.06538175 0.00
Total 0.94 334.38 304.36 334.38 358.95 326.19 358.95 469.50 556.85 469.50 307.09 278.42 307.09 87.35 225.20 272.54 0.00 0.00 0.00
Unit Cost [euro/kW] -0.06 3343.827385 1521.782667 3343.827385 3589.496336 1630.968867 3589.496336 4695.006614 2784.248108 4695.006614 3070.861884 1392.124054 3070.861884 2911.632009 2251.965382 2725.434584 1,364.83 777.95 1270.173618
Solar Field Cost [euro/m2] 280.3805229 159.6068996 280.3805229 280.3805229 159.6068996 280.3805229 269.3522391 148.1062813 269.3522391 292.8267966 166.1409738 292.8267966 329.1642804 162.2724846 309.6288109 0 0 0
O&M Cost [EuroM/year] 6.68765477 6.087130668 6.68765477 7.178992671 6.523875469 7.178992671 9.390013228 11.13699243 9.390013228 6.141723768 5.568496217 6.141723768 1.746979205 4.503930763 5.450869168 0 0 0
Electricity Produced [kWh] 219000000 438000000 219000000 219000000 438000000 219000000 306600000 876000000 306600000 219000000 438000000 219000000 70956000 455520000 219000000 65700000 227760000 219000000
Electricity Produced [kWh] using given LEC 1 218,465,414 423,632,933 218465413.8 242,773,537 467,581,199 242773537.3 319,796,180 906,446,327 319796180.1 216,859,050 417,815,104 216859050.4 71,093,830 480,623,910 236979431.9 66,650,465 226,409,611 19995139.56
218,465,413.75 423,632,932.76 218,465,413.75 242,773,537.26 467,581,198.99 242,773,537.26 319,796,180.13 906,446,326.90 319,796,180.13 216,859,050.42 417,815,103.81 216,859,050.42 71,093,829.56 480,623,910.26 236,979,431.87 66,650,465.21 226,409,610.62 19,995,139.56
O&M [R/kWh] 0.03 0.01 0.03 0.03 0.01 0.03 0.03 0.01 0.03 0.03 0.01 0.03 0.02 0.01 0.02 - - -
LEC [Euro/kWh] using given Enet -0.16 0.1819 0.0854 0.1819 0.1757 0.0829 0.1757 0.1745 0.0730 0.1745 0.1683 0.0792 0.1683 0.1460 0.0557 0.1367 0.0730 0.0408 0.059915316
LEC Rands 2.381279697 1.117743531 2.381279697 2.300283789 1.085345168 2.300283789 2.284084607 0.955751715 2.284084607 2.203088699 1.036747623 2.203088699 1.91150343 0.728963172 1.789263732 0.955751715 0.534572993 0.784459243
43780.06319 19924.3961 43780.06319 46996.55762 21353.94919 46996.55762 61470.78259 36453.60363 61470.78259 40206.18048 18226.80182 40206.18048 38121.41556 29484.53235 35683.56992 17869.41354 10185.56572 16630.12915
POWER TOWER DISH STIRLING DISH STIRLING SCOT CHIMNEY
Molten Salt PHOEBUS
ESKOM Solar Only Hybrid Solar Only Hybrid
Project near term mid term Long Term Long Term Scaled Mid Term Scaled near term mid term Scaled Long Term Scaled Short Term Long term Scaled Short Term Scaled Scaled Pilot Scaled
Plant Size [Mwe] 30 100 200 200 100 100 100 10 30 100 100 100 1 100 100 1 100 34 100 5 100
Solar Field [m2 x1000] 275 826 1490 2477 916.6666667 1350 1350 80.5 805 0 0 0 0
Adjusted solar Field 275.00 826.00 1,490.00 2,477.00 916.67 1,350.00 1,350.00 80.50 - 805.00 - - - - - - - - - - -
0 0 0 0
Thermal Storage[hrs] 6.5 6.5 6.5 13 6.5 13 13 1 3 10 8 0 0
Annual Solar CF 40% 41% 44% 74% 41% 19% 0.188 22% 40% 22% 38% 38% 24% 0.705882353 32% 6.40
Annual Solar/Elec % 0 0 0 -
Solar Fraction % 0 0 0 -
0 0 0 -
Capital Cost [euroM] 0 0 0 -
Infrastructure 4.78 8.19 13.65 13.65 14.80320401 10.92 10.91862003 0.00 0 4.09 68.30
Solar Field 45.04 110.41 142.35 236.66 141.2209454 180.16 180.1572305 51.45 143.2620553 20.06 344.60
Adjusted Solar Field Cost 45.04 110.41 142.35 236.66 141.22 180.16 180.16 - - - - - - - - - - 51.45 143.26 20.06 344.60
Tower/Receiver 24.57 34.12 51.86 68.24 75.27097012 47.77 47.76896264 23.88 65.13890146 15.42 250.10
HTF System 7.23 15.01 23.20 23.20 22.1631185 15.01 15.01310254 33.57 91.56668255 0.00 -
Storage 17.20 32.76 51.86 81.89 52.68967813 60.05 60.05241018 0.00 0 0.00 -
Adjusted Storage 17.20 32.76 51.86 81.89 52.69 60.05 60.05 - - - - - - - - - - - - - -
Power Block 38.22 77.80 113.28 113.28 117.0881732 38.22 38.21517011 5.60 15.26111376 8.46 137.22
Balance of Plant 4.78 8.19 13.65 12.83 14.80320401 10.92 10.91862003 0.00 0 0.00 -
Services 13.65 27.84 39.58 54.46 42.29486861 35.21 35.2125496 0.00 0 0.00 -
Land 0.00 0.00 0.00 0.00 0 0.00 0 0.00 0 0.00 -
Contingencies 22.66 45.99 65.38 88.58 70.2094819 58.14 58.14165167 12.97 35.72283072 1.50 25.04
Total 178.11 360.31 514.81 692.79 550.54 456.40 456.40 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 16.38 0.00 127.47 350.95 49.54 825.26
Unit Cost [euro/kW] 5936.999643 3603.144611 2574.064673 3463.932205 5505.436439 4563.983174 4563.983174 4,599.47 5,627.64 4008.758833 4,546.24 4008.758833 4,094.48 1,637.79 3110.30907 16,377.93 12441.23628 3749.261438 3509.515838 9908.647679 8252.647445
Solar Field Cost [euro/m2] 163.7793005 133.6737834 95.53792528 95.54343529 154.0592132 133.4498004 133.4498004 0.00 0 #DIV/0! #DIV/0! #DIV/0! #DIV/0!
0.00 0
O&M Cost [EuroM/year] 3.562199786 7.206289221 10.29625869 13.85572882 11.01087288 9.127966347 9.127966347 0 0 0 #REF! 0 0 0 0 0.00 0 2.549497778 7.019031676 0.990864768 16.50529489
Electricity Produced [kWh] 105120000 359160000 770880000 1296480000 359160000 164688000 164688000 19272000 105382800 192720000 335508000 335508000 0 0 0 0 0 71481600 618352941.2 14016000 5606400000
Electricity Produced [kWh] using given LEC 97,923,909 352,275,480 789,054,098 1,358,854,874 326,413,029.96 611,203,809 611203809 20,078,958 101,340,644 200789575.7 363,854,147 200789575.7 363854146.6 200789575.7 363854146.6 200789575.7 20078957573 81,618,679 240054937 14,869,302 297,386,044.47
97,923,908.99 352,275,479.68 789,054,097.68 1,358,854,873.55 326,413,029.96 611,203,809.00 611,203,809.00 20,078,957.57 101,340,643.81 200,789,575.73 363,854,146.61 200,789,575.73 363,854,146.61 200,789,575.73 363,854,146.61 200,789,575.73 200789575.7 81,618,678.58 240,054,937.00 14,869,302.22 297,386,044.47
O&M [R/kWh] 0.04 0.02 0.01 0.01 0.03 0.01 0.01 - - - #REF! - - - - - - 0.03 0.03 0.07 0.06
LEC [Euro/kWh] using given Enet 0.2161 0.1215 0.0775 0.0606 0.2004 0.0887 0.0887 0.2722 0.1980 0.186569232 0.1485 0.148471052 0.2475 0.0990 0.116252993 0.8649 0.406311893 0.1856 0.1737 0.3959 0.3298
LEC Rands 2.829754677 1.591287049 1.015060483 0.793186937 2.624058523 1.161735426 1.161735426 3.563819954 2.591869058 2.442713641 1.943901793 1.943901793 3.239836322 1.295934529 1.522077181 11.32344208 5.319760349 2.429877241 2.27449934 5.183738115 4.317396732
77731.94892 47175.25176 33701.71395 45352.57158 72081.57821 59755.31889 59755.31889 60219.92365 73681.54852 52485.87765 59523.01652 52485.87765 53608.24063 21443.29625 40722.65459 214432.9625 162890.6184 49088.33015 45949.38897 129731.9423 108050.2625

A Techno-Economic Feasibility Study On The Use of

Uploaded by

Copyright:

Available Formats

A Techno-Economic Feasibility Study On The Use of

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

A Techno-Economic Feasibility Study On The Use of

Uploaded by

Copyright:

Available Formats

A TECHNO-ECONOMIC FEASIBILITY STUDY ON THE

USE OF DISTRIBUTED CONCENTRATING SOLAR

Christiaan César Bode

A project report submitted to the Faculty of Engineering, University of

On this _________day of _________________________________ 2009

My supervisor, Professor T.J. Sheer, whose guidance is greatly appreciated.

Professor W. Cronje, of the School of Electrical and Information Engineering at the

Mr E. Brink, of the Electrical Engineering department at the University, for his

This study provides an evaluation of Concentrating Solar Power (CSP) technologies

1. Compound Linear Fresnel Receiver (CLFR) field with an Organic Rankine

TABLE OF CONTENTS vii

LIST OF TABLES xiii

LIST OF SYMBOLS AND ABBREVIATIONS xv

2.1 Utility and Distributed Generation 5

4.1 Economic Comparison of Existing Technologies 57

5.1 Wits Electricity Profiles 73

6.1 DNI Synthesis 94

7.1 Initial Design Results 99

8.1 Suitability of Design Approach 113

9.1 Conclusions 126

APPENDIX A STORAGE CONCEPTS 140

APPENDIX B TECHNOLOGIES USED IN THE COMPARISON 142

APPENDIX C SOLAR RADIATION DATA 150

APPENDIX D DATA VERIFICATION 153

APPENDIX F NATURAL GAS PRICING TARIFF FROM EGOLI GAS 155

APPENDIX G WITS UNIVERSITY USAGE AND BILLING TRENDS 156

APPENDIX H SPECIFIC COSTS 158

APPENDIX I PERSPECTIVE MODEL 159

APPENDIX J REFERENCE PLANT RESULTS 160

APPENDIX K MODEL INPUT 162

APPENDIX L COMPARISON MODEL 170

Table 2-1: Advantages and Disadvantages of CSP 24

ASf aperture area of the solar field m2

BOP balance of plant -

CLFR compound linear Fresnel reflector -

Esolar net solar electricity generated kWh

ECOSTAR European concentrated solar thermal road mapping -

kinsur annual insurance rate -

K fuel annual fuel costs R/$/€

KO&M annual operating and investment costs R/$/€

LEC Levelised Cost of Electricity (R/$/€)/kWh

PPA power purchase agreement -

Qɺ therm thermal energy input to the power cycle kW

Tamb ambient temperature K

TES thermal energy storage -

η piping piping efficiency -

η pbnet net power block efficiency -

ηopt optical efficiency -

ηrec / pip receiver/piping efficiency -

η s −e net solar to electric efficiency -

η Sf solar field efficiency -

η stor storage efficiency -

ξ geo geometric Efficiency -

ξ IAM the incident angle modifier -

ξS shading losses within the solar field -

ξcos cosine losses -

ρ reflectivity of the mirrors -

τ2 transmission factor of absorber tube -

As a result of insufficient electrical power generation infrastructure investment in

On this _day of _________________________ 2009

Esolar = Aa ⋅ DI a ⋅η s −e (1) (Pitz-Paal et al., 2005)