Techno-economic Appraisal of Concentrating Solar Power Systems (CSP)

(1)

Master Level Thesis

European Solar Engineering School No.169, August 2013

Techno-economic Appraisal of Concentrating Solar Power

Systems

Master thesis 18 hp, 2013 Solar Energy Engineering Student: Maria Gasti

Supervisor:

Tara Chandra Kandpal

Dalarna University Energy and Environmental

Technology

(2)

1

1 Introduction...2

1.1 Aims...2

1.2 Method...3

1.3 Related work...4

2 Simulations and Calculations...5

2.1 Boundary Conditions and Assumptions...6

2.2 Simulations...7

2.2.1 Parabolic Trough...9

2.2.2 Linear Fresnel Mirrors...17

2.2.3 Molten Salt Power Tower...23

2.2.4 Dish Stirling...27

2.2.5 Flat Plate PV...31

2.2.6 High-X Concentrating PV...35

3 Results...38

3.1 Parabolic Trough...39

3.2 Linear Fresnel Mirrors...42

3.3 Molten Salt Power Tower...45

3.4 Dish Stirling...48

3.5 Base Cases Comparison...51

3.6 Comparisons...54

4 Discussion and Conclusions...57

5 References...60

(3)

2

1 Introduction

The diffusion of Concentrating Solar Power (CSP) systems is currently taking place at a much slower pace than photovoltaic (PV) power systems. This is mainly because of the higher present cost of the solar thermal power plants, but also for the time that is needed in order to build them. Though economic attractiveness of different Concentrating technologies varies, still PV power dominates the market. The price of the CSP systems is expected to drop significantly in the near future and wide spread installation of them will follow. In this project it is proposed to examine the existing (CSP) systems in several different locations in Europe, including three different locations in Greece. An attempt to study the effect of various design, operation and economic parameters on the unit cost of electricity delivered by the CSP plants would be made in this project. Several feasible scenarios for operation of CSP systems shall be considered for analysis. Also a comparison of CSP systems with photovoltaic power plants shall be attempted. This will give an indication on how CSP can compete PV and maybe replace them in some cases. It has to do both with the technology and the financing of these systems. Last but not least, since CSP have the advantage to store power for many hours, the effect of design hours of storage on the unit cost of electricity would also be studied. So, an other question of this project is the definition of the more appropriate system for a particular geographical area. For the accomplishment of this work a literature study has been made. It includes the basic principles of the CSP systems and their possible future technological advancements (Lovegrove and Stein 2012).

1.1 Aims

The main aim of this project is the creation of different relevant case studies on solar thermal power generation and a comparison between them. The purpose of this detailed comparison is the techno- economic appraisal of a number of CSP systems and the understanding of their behaviour under various boundary conditions. The CSP Technologies which will be examined are the Parabolic Trough, the Molten Salt Power Tower, the Linear Fresnel Mirrors and the Dish Stirling. These systems will be appropriately sized and simulated. All of the simulations aim in the optimization of the particular system. This includes two main issues. The first is the achievement of the lowest possible levelized cost of electricity and the second is the maximization of the annual energy output (kWh). The input values that need to be changed and interpreted are many and a large number of combinations of them are possible, in case it gives better results. The project also aims in the specification of these factors which affect more the results and more specifically, in what they contribute to the cost reduction or the power generation. Also, a photovoltaic (PV) system will be simulated under the same boundary conditions to facilitate a comparison between the PV and the CSP systems. Depending on the boundary conditions, the selection of the system which performs better can be different.

(4)

3

1.2 Method

The project is carried out with the use of the System Advisor Model (SAM) software of National Renewable Energy Laboratory (NREL). This tool considers for each type of technology various input parameters, which determine the system’s performance. A brief description of SAM and its input parameters is presented in the following paragraphs (SAM Help Contents).

To begin with, the most important parameter SAM takes into account is the climate. It refers to the location of the plant. Typical meteorological data for a certain area are taken from NREL weather database (Energy Plus weather data 2011). They include information about the latitude, longitude and the elevation of the particular location. Also, they include hourly data for direct normal and global horizontal radiation, dry bulb temperature and wind speed. It is worth noting that each file contains 8760 values for an entire year.

The next parameter is about the financing option of the examined CSP system. Utility Independent Power Producer is defined for the particular project, which refers to a seller who purchases electricity according to a power purchase agreement between two parties (seller-buyer). The financing option requires the following values to be determined from the SAM user. The installation cost and the operation-and-maintenance cost. More specifically, the total installed cost consists of the direct and the indirect capital cost. The direct one is related with the premeditated expenses for the completion of the project, like the solar field’s cost, the pumps and the piping and other various expenses which have already been defined. The Indirect one includes design and construction expenses, like permitting, engineering project and development activities and other issues which affect the project as an entity. It is also dependent on the land area and on the amount of the direct cost which has been subjected to sale taxes.

SAM considers financial and incentives assumptions too. The principal values for these assumptions are those of the inflation rate, the discount rate and the life time of the plant. Furthermore, the construction period and the loan parameters are taken into account in a long range planning. Here, the power purchase agreement price per kWh and its escalation rate is defined too. Regarding the type of investment, incentives can be considered if the project is based on them.

In addition, SAM assumes design parameters, such as the solar field, the power block, the solar collector assembly and the thermal storage. A short report for each of the aforementioned components of the plant is following. Important information regarding the solar field is the field’s layout, the solar multiple and the type and temperature of heat transfer fluid. The power block field includes values of the estimated net output after its reduction due to parasitic losses. The solar collector assembly refers to the physical properties of the CSP technology. This is about its dimensions, its optical characteristics and the relative heat loss parameters. The thermal energy storage indicates the thermal storage capacity that the power block delivers. This is described on number of hours. A more detailed explanation on the storage inputs concerns the type of the storage fluid and the limits of the amount of the storage energy.

As it was described in the previous lines, SAM software takes a great variety of input values. Due to this large amount of input data, some of them were not presented. It is worth noting to mention that SAM software contains defaults values. The project’s strategy is the alteration of one default input value in every simulation of CSP system and the detailed interpreting of its influence on the results. A combination of input values can be possible after that as well.

More specifically the procedure consists of the next steps. The first step is the variation of the climate parameter. The selected locations belong to Greece, Spain and Italy. The Greek locations are Athens, Thessaloniki and Andravida. Between them the climate is quite the same, as Greece is not a big country. The main characteristic of these regions is the intensity of direct radiation which is very high and the land availability, which makes them suitable for such installations. Weather data will be compared from different sources for these regions, but since the accuracy and the availability of

(5)

4 them is insufficient, other locations will replace them. Also, a comparison in the results will be made between them and the Italian and Spanish location. The next step is the solar multiple and thermal storage energy variation. The last part of simulations includes potential improvements on the existing CSP systems and technological advancements. According to (Hank Price et al 2002), a higher reflectivity could be examined. But SAM already considers high reflectivity, which exceed the 0.9 and 0.93 in some cases. The CSP systems’ study will be continued with a detailed comparison between them. Then, a photovoltaic (PV) system will be simulated under the same boundary conditions and be compared with the relative CSP study cases.

The drawback of the SAM software is its limitation in modelling new CSP systems. Even small changes in the geometries and the systems component s combination is difficult and maybe impossible in some cases to be modelled. SAM uses only its existing models for the prediction of the results and rejects the option of simulating new CSP designs (Wagner and Zhu 2011).

1.3 Related work

In this paragraph a brief description of the literature study will be made. Except from the literature which concerns the System Advisor Model software a further study in the CSP systems was required.

Principles and basic components of the CSP systems were provided from the (Lovegrove and Stein 2012) book. This book and (Hank Price et al 2002) includes future trends and technological advancements in the CSP systems. Finally, it provides information about various meteorological and climatic parameters which influence the CSP systems’ performance. More information about the Dish Stirling and its operation were taken from (Fraser 2008), where its components were studied in order to determine and to create a Dish Stirling prediction model. The report of (Yinghao Chu 2011) gives more information about the CSP and PV mechanisms. It also evaluates their performance and compares real CSP and PV systems according to their penetration on the market, their outputs and their costs. (Panos and Kretschmann 2010) is a relevant project where there are detailed comparisons of existing CSP systems and possible variations in their thermal storage capacity and their solar multiple. Last but not least, (NASA) website was used for the selection of further weather data.

(6)

5

2. Simulations and Calculations

The simulation results which will be analysed are the levelized cost of electricity (LCOE) and annual electric output. The LCOE in other words is the unit cost of electricity which should be sold in order to cover the installed cost, the operation and maintenance costs and all the financing. SAM provides two kinds of LCOE, The real and the nominal one. Nominal LCOE refers to current value of money.

Real LCOE considers inflation during the life time of the project in contrast with the nominal one which excludes it from its calculations. But the main difference among them relies on the analysis which will be made. The real LCOE is mainly used for the long term projects, rather than the nominal, which is suitable for short term schemes. In this project LCOE (real-w/o incentives) is considered for the comparison of the variants, in which tax credits and incentives payment is excluded. The annual electric output is the total annual output for the first year of the project’s cash flows. The thesis first presents the boundary conditions and the assumptions that have been considered in the study followed the simulations made in the study. Finally the results of the simulation studies are used to compare different scenarios of concentrator design and operational conditions.

(7)

6

2.1 Boundary Conditions and Assumptions

The boundary conditions refer to all the CSP system models which are going to be simulated. In this way the comparison between them is reasonable and feasible to be made. Some of them concern the financing and some others the technology. Here is a list of these assumptions.

 A utility independent power producer is considered for the financing option. The CSP system is owned and operated by a producer who sells the produced energy at a certain price. This price is defined by the power purchase agreement (PPA). The project is performed from the owner’s and producer’s viewpoint.

 Useful life of CSP systems is assumed to be 30 years

 Nominal power of the system is 100MW. In SAM software it is mentioned as the estimated net output at design and describes the net output after the reduction due to the parasitic losses.

 The financing mode of the system works according to a specified internal rate of return (IRR) target of 15%, which is the minimum required value. The IRR refers to the return on investment (the profits). In order to meet this target IRR value, SAM software calculates the PPA price, which is the defined amount of money the buyer pays for each kWh. It is worth noting that the PPA value will have an escalation rate of 1.2% per year.

 A construction loan which covers the total installed cost of the CSP system. This applies to all the case studies. For example, when the solar multiple raises and the installed cost increases then, the loan also covers this capital cost.

 Inflation rate of 2.5%

 Solar multiple of 1 in the basic cases

 Thermal storage capacity of 6 hours for the basic cases which have thermal storage

(8)

7

2.2 Simulations

Figure 1 illustrates the solar irradiation map of Europe (GeoModel Solar 2011). This project focuses on the Mediterranean countries and more specifically in the cities which are considered in the study.

The selection of these countries is due to their location and their availability of direct normal irradiation. In the first place the project is supposed to focus on Greece, but there is the limitation of the availability of weather data. So for Greece, data for three cities are only available. Moreover, four Spanish and three Italian cities will be simulated.

Figure 1. Solar irradiation map of Europe

Solar recourses are enough in all of the considered cities, but each of them has different meteorological parameters which influence the power output. Here is a brief description of their importance and influence in the CSP system performance according to (Lovegrove and Stein 2012).

The primary parameter of a location with the greatest significance on a CSP plant is the direct normal irradiance (DNI). It determines the energy production capability (energy yield) and according to this value a location can be considered suitable or not for a CSP plant. The outputs are predicted according it and the rest of the parameters, which will be discussed later, can affect the result, but they are not as extremely significant as DNI.

The air temperature is described by the dry bulb temperature in Table 1. It can affect both the solar field’s and the power block’s efficiency. The effect of the ambient temperature is the convection losses of the heat transfer fluid and the collector’s area (reflectors). For the power block efficiency, humidity and ambient pressure have to be taken into account too. The elevation of each location

Thessaloniki

Athens Andravida

Brindisi Palermo

Malaga Cagliari Sevilla

Toledo

Alicante

(9)

8 determines the ambient pressure. Also, relative humidity is considered in the weather data of each location, but there is not an annual mean value for it. Increased humidity and low air temperature reduces power block’s efficiency and therefore the overall energy output of the system.

High wind speeds have serious impact in the CSP system’s performance. This mainly happens for two reasons. The first is that they create vibrations and in some cases they may bend the mirrors. Then the focusing changes and power generation goes markedly down. So, construction is required for the resistance in the strong wind and the maintenance of the appropriate geometry. The second reason is that wind causes convective heat losses in the solar field.

Several characteristics of each location considered in the study are presented in Table 1. Every place combines different meteorological parameters, but in all of them DNI is plenty, which makes them suitable for studying. The simulation of them will determine which parameters have the greater effect on a CSP system and how they can be managed. The displayed numbers that refer to weather are the average annual values.

Table 1. Locations’ climate characteristics

Location DNI Wind Speed Dry Bulb Temperature Latitude Elevation

Units (kWh/m2) (m/s) (C°) (Degrees) (m)

Athens 1519.8 3.2 17.9 37.90 15

Thessaloniki 1372.7 3.1 15.4 40.52 4

Andravida 1151.7 2.8 16.7 37.92 12

Malaga 1988.1 6.7 18.0 36.67 7

Sevilla 1772.7 2.7 18.4 37.42 31

Toledo 1926.9 6.7 15.6 39.88 529

Alicante 2016.8 6.7 17.9 38.37 31

Cagliari 1013.1 4.0 16.4 39.25 18

Palermo 1516.9 3.8 18.8 38.18 34

Brindisi 1119.9 5.1 16.6 40.65 10

After the climate change, the solar multiple will be examined. A solar multiple of 1, expresses the required solar field aperture area to meet the power cycle demand. More specifically, it expresses the amount of thermal energy needed to drive the power cycle at design conditions. It is interesting to study the increase of this parameter and optimize the solar field size for a given power cycle capacity, location and thermal energy storage.

Thermal Storage Capacity is considered to be one of the advantages of the CSP systems. Its importance relies on the fact that it can shift the electricity use from high consumption periods to less expensive ones, by saving money and improving the overall system efficiency as well. The effect of this parameter is dependent from the power cycle thermal input. The simulations include a case without Thermal Storage and three others with gradual increase until 12 h of thermal storage.

(10)

9

2.2.1 Parabolic Trough

The simulations presented in this section will be used as the bases cases for each location for Parabolic Trough model. They will be used as references to be compared with different system variations. Afterwards, a comparison between the location and their possible variants will be done.

The plant corresponds to a capacity of 100MW.

Figure 2 shows the estimated values of LCOE for three cities of Greece. Athens has the minimum LCOE while its value is highest for Andravida. The reason for the same may be noted from Figure 3 as the estimated annual electricity output is highest for Athens. Similar values of LCOE and annual electricity output for four cities of Spain are presented in Figures 4-5 and for three cities in Italy in Figures 6-7.

Figure 2. Base Case-LCOE for Greece Figure 3. Base Case-Annual Output for Greece

The base cases for the Spanish cities in Figure 4 bring much better results than in Greece. The LCOE for the four climates is low, with Malaga having the lowest value. Similar LCOE values have Toledo and Alicante and Sevilla has just a little bit higher price of electricity. The power outputs are higher for the locations with lower LCOE.

Figure 4. Base Case-LCOE for Spain Figure 5. Base Case-Annual Output for Spain

The Italian cities except from Palermo do not have such a good performance. Cagliari has the worse LCOE which reaches 122.95 Cents/kWh and Brindisi has very high LCOE as well. Palermo’s performance is better and its electric output is the highest of the Italian locations.

65.42

79.92 90.69

0 20 40 60 80 100

Athens Thessaloniki Andravida LCOE(real without incentives)

92.39

75.56

66.56

0 20 40 60 80 100

Athens Thessaloniki Andravida Annual Output (GWh)

49.57

56.25

51.90

50.23

46 48 50 52 54 56 58

Malaga Sevilla Toledo Alicante LCOE(real without incentives)

122.10

107.53

116.58

120.49

100 105 110 115 120 125

Malaga Sevilla Toledo Alicante Annual Output (GWh)

Cents/kWh GWh

(11)

10 Figure 6. Base Case-LCOE for Italy Figure 7. Base Case-Annual Output for Italy

The next set of simulations is about the variation of the solar multiple. Figures 8-13 show the variations in LCOE and annual output with changes in the values of the solar multiple.

To begin with, Athens has its minimum LCOE for 2.5 solar multiple. Here further increase or decrease does not produce the sufficient energy to minimize the LCOE and cover the initial cost. Thessaloniki has its minimum for solar multiple of 3 and Andravida for solar multiple of 2.5. In Figure 9, it is noticeable that the power output has a smaller increase between 2.5 and 3 solar multiple than between 2 and 2.5.

Figure 8. Solar Multiple-LCOE for Greece Figure 9. Solar Multiple-Annual Output for Greece

In Figure 10 all the Spanish climates have their power output for a 2.5 solar multiple. In Sevilla LCOE fluctuates in an interesting way, as solar multiple changes. For solar multiple of 2 Sevilla gets a high LCOE value equal to 35Cents/kWh. Then in solar multiple of 2.5, there is a sharp reduction in the price of electricity and it becomes equal to that of the Toledo. Finally, there is an increase again for a solar multiple of 3. The rest of the climates behave in the same way, but with smoother fluctuations.

122.95

66.12

105.34

0 20 40 60 80 100 120 140

Cagliari Palermo Brindisi LCOE(real without incentives)

49.05

91.40

57.27

0 20 40 60 80 100

Cagliari Palermo Brindisi Annual Output (GWh)

40.4 38.0 38.7

48.3

44.0 43.8

55.1 51.8 52.7

0 10 20 30 40 50 60

2 2.5 3 3.5

Athens Thessaloniki Andravida

205.4

248.2 272.7 171.6

213.9 240.7 150.4

181.7 199.9

0 50 100 150 200 250 300

2 2.5 3 3.5

Athens Thessaloniki Andravida Cents/kW

h

Solar Multiple

GWh

Solar Multiple GWh

Cents/kWh

(12)

11 Figure 10. Solar Multiple-LCOE for Spain Figure 11. Solar Multiple-Annual Output for Spain

Figure 12 presents the solar multiple variations for Italy. Cagliari and Brindisi reach their minimum LCOE for solar multiple of 3 and the related difference in the LCOE is big. On the contrary, Palermo has its minimum LCOE for solar multiple of 2.5. It is seen that the climates which do not have a high potential for a good CSP system performance reach their minimum LCOE for a solar multiple of 3 and not for 2.5 like the climates with the higher electric outputs. Moreover, the solar multiple changes cause them significant decrease in the LCOE, which does not happen in the other locations where the decrease is very small compare to them.

Figure 12. Solar Multiple-LCOE for Italy Figure 13. Solar Multiple -Annual Output for Italy 31.4

30.8 31.5 35.0

32.7

33.1 33.1 32.7

33.9

31.7

31.1

31.8

30.5 31 31.5 32 32.5 33 33.5 34 34.5 35 35.5

2 2.5 3 3.5

Malaga Sevilla Toledo Alicante

237.8

288.4

318.8

251.0

288.6 311.7 262.1

304.2

332.6

265.4

307.0

335.2

225 245 265 285 305 325 345

2 2.5 3 3.5

Sevilla Toledo Alicante Malaga

71.7

63.7 59.9

40.9 38.6 39.1

62.9

56.4 55.0

0 10 20 30 40 50 60 70 80

2 2.5 3 3.5

Cagliari Palermo Brindisi

115.4

147.7 175.6 202.9

244.5 270.2

131.7

166.7 191.4

0 50 100 150 200 250 300

2 2.5 3 3.5

Cagliari Palermo Brindisi Solar Multiple

Solar Multiple

Solar Multiple Solar

Multiple

Cents/kWh GWh

GWh Cents/kWh

(13)

12 Thermal Storage Capacity is the next variant. The examined capacity is between 0 and 12 hours for S.M of 1.

Athens appears to have almost the same electric output for 4h, 6h and 12h storage capacity. But a great difference in the LCOE is seen between them. So the 4h is the best thermal storage as it combines the same output in the lower price of electricity. The same thing happens with Thessaloniki and Andravida as well. And the best combination for them is the 4h storage.

Figure 14. Storage Capacity-LCOE and Annual Output for Greece-S.M:1

In Spain the electric outputs are very high and they exceed the 100GWh. Similar to the Greek climates, in Spanish cities the best case is the one of the 4h storage capacity, where the LCOE is very close to the case without storage and the power output is annual output similar to the one of the 12h thermal storage.

Figure 15. Storage Capacity-LCOE and Annual Output for Spain-S.M:1

In Cagliari and Brindisi the results are disappointing, as the LCOE are very high. Palermo case has much better results, close to those of Athens. Again the best combination for the three locations is the one for thermal storage of 4h.

0 20 40 60 80 100 120

140 LCOE (real-w/o

incentives) Annual Output (GWh)

0 20 40 60 80 100 120

140 LCOE (real-w/o

(14)

13 Figure 16. Storage Capacity-LCOE and Annual Output for Italy-S.M:1

In the Solar Multiple variations, it is noticeable that in most of the cases the LCOE (real without incentives) takes its minimum value when S.M is 2.5. In some others the unit cost of electricity is the optimum for S.M of 3, but with very small difference than in the price that it gets for S.M of 2. So, for optimizing the size of the field it is interesting to simulate the Storage Capacity for a S.M of 2.5. In these graphs LCOE real is also presented, in which the unit cost of energy considers the incentives.

Figure 17 presents the thermal storage capacity variations for the parabolic trough model in Greece.

For Athens and Andravida, solar multiple of 2.5 was the optimum and the results for the electric output become much higher with the storage variations. Athens combines the lower LCOE with and without incentives and the better annual output. The prices of electricity remains almost the same, but the output makes the difference. Thessaloniki and Andravida have similar behaviour. The optimum thermal storage capacity for Athens and Andravida is 12h and Thessaloniki is 6h, but the results with 12 h are very close.

Figure 17. Storage Capacity -LCOE and Annual Output for Greece-S.M:2.5 0

20 40 60 80 100 120 140 160

180 LCOE (real-w/o

0 50 100 150 200 250

300 LCOE (real-w/o

incentives) Annual Output (GWh) LCOE real

(15)

14 In Figure 18, there is the performance of the parabolic trough system in the four Spanish locations.

The power outputs for 12h exceed the 300GWh and the LCOE remain in very low levels. All the cases have their best combination for this storage capacity.

Figure 18. Storage Capacity-LCOE and Annual Output for Spain-S.M:2.5

In Figure 19, the best combination of solar multiple and thermal storage for Palermo is in 12h. In Cagliari and Brindisi the optimum solar multiple was for 3 and not for 2.5. The expected results are not supposed to be good and the best case for them is this of the 6h storage.

Figure 19. Storage Capacity-LCOE and Annual Output for Italy-S.M:2.5 0

50 100 150 200 250 300 350

400 LCOE (real-w/o

0 50 100 150 200 250

300 LCOE (real-w/o

(16)

15 The Installation cost for the previous cases is displayed in Figures 20-22. This cost differs from case to case, but remains the same between the countries (locations).

Base Case

The Installation cost is US $ 514.53 million for all the locations.

Solar Multiple Variations (Thermal Storage Capacity: 6h)

The solar multiple variations have great influence on the capital cost. From solar multiple of to 1 to 3, the installation cost is almost the double. The other solar multiple values correspond to high initial costs too.

Figure 20. Installation cost for different values of Solar Multiple (S.M)

Thermal Storage Capacity for S.M:1

In the thermal storage capacity variations, the initial cost does not increase as much as in solar multiple. For example in Figure 21 the increase between the storage of 6h and 12h is 173.72 million$

and the capital cost does not exceed the 700 million$.

Figure 21. Installation cost for Storage Capacity variation 514.53

726.96 833.17

939.38

0 200 400 600 800 1000

1 2 2.5 3

340.82 456.63

514.53

688.25

0 200 400 600 800

0h 4h 6h 12h

Solar multiple

Million $

Installation Cost

Million $

(17)

16 Thermal Storage Capacity for S.M:2.5

Figure 22 refers to the intial cost of the combinations between the 2.5 solar multiple and the thermal storages. Here the capital costs are the higher as they correspond to better results.

775.27 833.17

1006.89

0 200 400 600 800 1000 1200

0h 4h 6h 12h

Million $

Installation Cost

(18)

17

2.2.2 Linear Fresnel Mirrors

The following graphs are the base cases for the Linear Fresnel Mirrors model. The next simulations are for 100MW Nominal Power.

In Figure 23, the LCOE for the three Greek cities is presented and in Figure 24 are the related annual power outputs. The LCOE for the base cases are lower than in parabolic trough and the annual outputs as well. Athens has the best performance, the Thessaloniki and in the end Andravida. The power outputs are relatively low even for Athens which is the most suitable Greek location for CSP systems.

Figure 23. Base Case-LCOE for Greece Figure 24. Base Case-Annual Output for Greece

The LCOE in Figure 25 for the Spanish cities is lower than in the base cases for parabolic trough model. Again the annual output is not so good. Moreover in the previous simulations Malaga’s and Alicicante’s results were almost similar. But in Linear Fresnel Mirrors, Malaga performs much better.

Also in Sevilla is the less efficient system, but the difference in the LCOE between the four locations is not as big as in parabolic trough’s simulations, due to the low power outputs and the less efficient system.

Figure 25. Base Case-LCOE for Spain Figure 26. Base Case-Annual Output for Spain 54.2

67.7 75.4

0 10 20 30 40 50 60 70 80

Athens Thessaloniki Andravida LCOE (real-w/o incentives)

66.3

53.0

47.5

0 10 20 30 40 50 60 70

42.9

46.3

45.1 45.0

41 42 43 44 45 46 47

Malaga Sevilla Toledo Alicante LCOE (real-w/o incentives)

83.8

77.6

79.7 79.9

72 75 78 81 84 87

Cents/kWh GWh

(19)

18 In Figure 27, Palermo reaches a relatively low LCOE and the other two Italian cities have high unit costs of electricity. The annual outputs are low, like in the other countries. Palermo can be characterized a bit worse in its performance compared to Athens. Cagliari and Brindisi do not have any other limits for lower power outputs as their results are already bad. In this case Linear Fresnel Mirrors is not a good choice for these locations.

Figure 27. Base Case-LCOE for Italy Figure 28. Base Case-Annual Output for Italy

The presentation of results for Linear Fresnel Mirrors begins with the effect on LCOE of a variation in the Solar Multiple. The results for LCOE (real without incentives) and Annual Output are in the next graphs.

In Figure 28, the minimum LCOE is succeeded in the case with a solar multiple of 2 for all the Greek locations. But the difference with the electricity costs of the other solar multiples is not big. In Figure 29 is the annual output which increases sharply with the solar multiple changes, but still the results are not as good as in parabolic trough.

Figure 29. Solar Multiple-LCOE for Greece Figure 30. Solar Multiple-Annual Output for Greece

109.3

55.0

90.0

0 20 40 60 80 100 120

Cagliari Palermo Brindisi LCOE (real-w/o incentives)

32.7

65.2

39.8

0 10 20 30 40 50 60 70

35.9 37.1

39.3 41.4 41.7 43.5

48.6 49.7

51.8

30 35 40 45 50 55

1.5 2 2.5 3 3.5

149.9

168.9

181.6

130.0

150.1

163.8

110.7

125.9

137.5

100 110 120 130 140 150 160 170 180

1.5 2 2.5 3 3.5

Athens Thessaloniki Andravida Cents/kWh

Solar Multiple

Cents/kWh GWh

Solar Multiple GWh

(20)

19 In all the Spanish climates LCOE fluctuates in the same way as in Greece (Figure 31). The minimum value it gets is for solar multiple of 2 and between 2 and 3 the difference in price of electricity is almost 2 cents/kWh. In Figure 32, the electric output increase is great but no value exceeds the 230GWh. It is also significant to mention that Toledo has the worst performance and not Sevilla, as in the rest of the simulations. The best case will be determined from the lowest LCOE and this is for solar multiple of 2.

Figure 31. Solar Multiple-LCOE for Spain Figure 32. Solar Multiple-Annual Output for Spain

Here it is obvious that Palermo’s performance is outstanding compared to the other climates.

Cagliari and Brindisi have their minimum LCOE in solar multiple of 2.5, while Palermo’s minimum is in 2. In Figure 34 the greater increase in the annual output is seen for Palermo, then in Brindisi and finally in Cagliari.

Figure 33. Solar Multiple-LCOE for Italy Figure 34. Solar Multiple- Annual Output for Italy 29.8

30.6

32.1 31.1

31.9

33.8

31.6

32.6

34.3

30.4

31.1

32.6

29 30 31 32 33 34 35

1.5 2 2.5 3 3.5

180.8

204.8

222.4

173.2

196.3

211.7

170.5

192.5

208.3

177.2

201.8

219.0

165 175 185 195 205 215 225

1.5 2 2.5 3 3.5

57.5

55.1 55.8

36.6 37.6 39.7 52.5 52.0 53.5

30 35 40 45 50 55 60

1.5 2 2.5 3 3.5

93.3

113.4

127.7 147.3

166.6

179.7

102.3

120.2

133.3

85 95 105 115 125 135 145 155 165 175 185

1.5 2 2.5 3 3.5

Cagliari Palermo Brindisi Solar

Multiple Cents/kWh

Solar Multiple GWh

Solar Multiple Cents/kWh

Solar Multiple GWh

(21)

20 The power cycle always operates at its design point, but the produced energy from the solar field is not always enough for these standards. For this reason there is a backup system which replaces the amount of the missing energy. The fossil fill fraction refers to this amount of energy which fills the boiler.

For Solar multiple of 1, a percentage of auxiliary fractions from 0 to 75% will be tried in the system. In SAM it is called Fossil Fill Fraction and operates as a backup system, when the delivered solar energy is not enough to drive the power cycle at full load. This is the case of auxiliary fraction equal to zero. The auxiliary energy comes from a burner boiler and when the delivered solar energy drops below a certain fraction (which will be studied), the boiler supplies the rest of the required energy to reach the desirable.

The most interesting case is the one with the fossil fraction of 0.25, because fractions of 0.5 and 0.75 are very high to be used. The purpose of the project is to use the fossil fuel as a backup system and not as the main which will produce the most of the energy. So, the case of 0.25 fossil fuel fraction is the one which will be discussed. Between 0 and 0.25 fossil fraction, there is a great difference in the annual output. In all the Greek cities it exceeds the 150GWh. More specifically, in Athens it reaches 200GWh and in the other two climates it is very close to this value.

Figure 35. Fossil Fraction-LCOE and Annual Output for Greece

In Figure 36 is the hybridization of the Linear Fresnel Mirrors system with fossil fuel fraction for Spain. All the Spanish locations exceed 200GWh power output in the case of 0.25 fossil fractions.

For higher percentage of fossil fractions, the annual outputs become almost the same, as the generated energy from the fossil fuel sources are much more than the energy of the CSP system.

Figure 36. Fossil Fraction-LCOE and Annual Output for Spain 0

100 200 300 400 500 600

700 LCOE (real-w/o

0 100 200 300 400 500 600

700 LCOE (real-w/o

(22)

21 After the fossil fuel source integration in the hybrid system, the LCOE in Brindisi and Cagliari decrease significantly. It is noticeable that for 0.25 and for higher percentage of fossil fraction the annual outputs become the same for all the cases.

Figure 37. Fossil Fraction-LCOE and Annual Output for Italy 0

100 200 300 400 500 600

700 LCOE (real-w/o

(23)

22 The determination of the installed cost for the case studies follows:

Base Case

The Installation cost is US $ 281.30 million for the 100MW nominal capacity plants for all the locations. Compared to the parabolic trough installation cost for the base case, which is US $ 514.53 million, it is almost the half. But the results in the annual outputs are not as good as in the previous CSP system. Parabolic trough succeeded its maximum annual output for Malaga with 122.1 GWh and the maximum power output for Linear Fresnel Mirrors is for Malaga again and corresponds to 83.8 GWh.

Solar Multiple Variations

The solar multiple is the factor that causes the higher expenses in a plant. In the Linear Fresnel Mirrors case the cost is relatively low and even the case of solar multiple of 2.5 has similar cost to the parabolic trough base case.

Figure 38. Installation cost for different values of Solar Multiple (S.M)

Fossil Fill Fraction

The total capital cost is not affected from the variation of the fossil fill fraction and remains constant.

It is the same as the base case (281.30 million$). This is because SAM does not consider the fossil backup system in its financing and gives 0$/kWe for it in the direct capital cost. But the fossil fuel cost will increase the operation cost during project’s lifetime.

450.27 533.32

616.37

0 200 400 600 800

2 2.5 3

Solar multiple

Installation Cost

Million $

(24)

23

2.2.3 Molten Salt Power Tower

Here are the Base Cases for the Molten Salt Power Tower Model. Like the previous CSP systems, Molten Salt Power Tower plant has 100MW capacity. In the Molten Salt Power Tower, the Solar Multiple variations are not simulated. This is because it does not consider the heliostat field.

SAM software provides a fixed Power Tower model, where the S.M is 1.9. For a possible change of that, the solar field layout (the heliostat scale) remains the same, as the software does not provide another layout. This is one of the SAM’s limitations as it was mentioned in the methods section. For this reason this CSP system will be simulated with S.M of 1.9.

Figure 39 shows the unit cost of electricity and Figure 40 presents the annual power outputs for the Power Tower system in Greece. Athens has the best performance, reaching 227.3GWh and the lowest electricity price. The second location is Thessaloniki, where the results are close to these of Athens and last is Andravida, which has the lower annual output. The power outputs are much higher than these of Parabolic Trough and Linear Fresnel Mirrors, because of the different solar multiple.

Figure 39. Base Case-LCOE for Greece Figure 40. Base Case- Annual Output for Greece

Molten Salt power Tower results are very good in Spanish cities. The minimum power output is the one of Sevilla and the maximum this of Alicante. The LCOE fluctuate from 24 to 25.9 cents/kWh which is the lowest price of electricity so far.

Figure 41. Base Case-LCOE for Spain Figure 42. Base Case- Annual Output for Spain

31.2 35.5

43.7

0 10 20 30 40 50

227.3

199.8

161.9

0 50 100 150 200 250

24.1

25.9

24.9

24.0

23 23.5 24 24.5 25 25.5 26 26.5

295.1

274.0

285.3

296.1

260 270 280 290 300

Malaga Sevilla Toledo Alicante Annual Output (GWh) Cents/kWh

Cents/kWh

GWh

(25)

24 In Figure 43, Palermo has almost the same LCOE with Athens and its annual output is very close to the 227.3GWh of Athens. Cagliari and Brindisi take their lowest values of LCOE in this CSP system.

Comparing them to the other base cases the LCOE are not so bad. But since power tower refers to a solar multiple of 1.9, these prices are high.

Figure 43. Base Case-LCOE for Italy Figure 44. Base Case- Annual Output for Italy

The results for the Thermal Storage Capacity variation are shown in the next charts.

In Figure 45 it is obvious that the cases without any storage have negative effects in the annual outputs. In the rest of the cases, the annual output is much higher but does not have great difference in GWh between 4h, 6h and 12h storage. In Athens, the LCOE remains the same between 4h and 6h. There is an increase in LCOE for the 12h storage, since this size is too big for the given size of the solar field. In Thessaloniki and Andravida, it happens the same for the 12h storage.

Figure 45. Storage Capacity-LCOE and Annual Output for Greece 60.4

31.7

51.7

0 10 20 30 40 50 60 70

117.0

223.7

136.7

0 50 100 150 200 250

38 31 31 34 41 35 35 38 51 43 44 47 26 21 22 23 29 24 25 27 35 30 30 33 0

50 100 150 200 250

LCOE (real-w/o incentives) Annual Output (GWh) LCOE real

Cents/kWh GWh

(26)

25 In Figure 46, the Spanish locations have more extreme increase in their power outputs as the storage capacity become larger. In Malaga, Toledo and Alicante the annual output increases rapidly from 4h to 6h storage. On the contrary, this does not happen to Sevilla and in Greece, which had their maximum output for 6h storage but with very small difference compare to 4h and 12h. Also, it is noticeable that in Toledo the LCOE reach its minimum in the case for 6h thermal storage.

Figure 46. Storage Capacity-LCOE and Annual Output for Spain

In Figure 47, the annual outputs for Palermo are far better than those of the other two Italian cities and they exceed 200GWh. Cagliari and Brindisi do not reach even 150GWh, which proves their bad performance. The LCOE of Palermo is equal to this of Athens for 4h. But this value does not correspond to the optimum case, which is the one of the 6h storage and it has the highest output.

Figure 47. Storage Capacity-LCOE and Annual Output for Italy 30 24 24 26 32 26 26 28 31 26 25 27 30 24 24 26

21 17 17 18 22 18 18 19 22 18 17 19 21 17 17 18 0

50 100 150 200 250 300

350 LCOE (real-w/o

62 59 60 65

38 31 32 34 57 50 52 56 43 41 42 45

27 22 22 24 40 35 36 39 0

50 100 150 200

250 LCOE (real-w/o

(27)

26 The determination of the installed cost for the case studies follows:

Base Case

The Installation cost is US $ 610.95 million for all the locations. As it was mentioned before, this price corresponds to a solar multiple of 1.9. The previous CSP systems of solar multiple of 2 had the following installation costs. Linear Fresnel Mirrors had US $ 450.27 million and Parabolic Trough had US $ 726.96. The Power Tower has lower installation cost and it provides better results than Parabolic Trough.

Thermal Storage Capacity Variation

The installation cost is not as high as it was expected when the thermal storage capacity becomes the double, from 6h to 12h. In the previous CSP systems, the increased amount of money was greater

592.97 610.95

664.89

500 550 600 650 700

0h 4h 6h 12h

Installation Cost

Million $

(28)

27

2.2.4 Dish Stirling

In the Dish Stirling model, the variation of the Solar Field size and the layout of the network will be expressed in Solar Multiple which corresponds to different capacity. Dish Stirling consists of a Stirling Engine, which is a closed-cycle engine and operates individually. The unit nameplate capacity of each Dish is 25 kW. There will be a variation in the number of collectors and in the total solar field area.

The primary number of collectors is 4000 and corresponds to a total capacity of 100MW. For these values the Base Cases are the followings.

The base cases for Greece are presented in Figure 49 for the LCOE and in Figure 50 for the annual output. The produced energy for Athens is 113.4GWh, which is the highest so far for solar multiple of 1. Also Thessaloniki and even Andravida have reached very high power outputs.

Figure 49. Base Case-LCOE for Greece Figure 50. Base Case- Annual Output for Greece

In Figure 51 the Spanish LCOE are just a little bit higher than these of Molten Salt Power Tower.

Moreover, the power outputs in Figure 52 are very high, with Toledo having the lowest value this time instead of Sevilla. Also, Malaga is the most efficient location for the Dish Stirling system.

Figure 51. Base Case-LCOE for Spain Figure 52. Base Case- Annual Output for Spain

31.1 35.3 39.6

0 10 20 30 40 50

113.4

99.6

81.4

0 20 40 60 80 100 120

25.2

25.7

26.1

24.9

24 24.5 25 25.5 26 26.5

140.1

137.4

135.3

141.9

132 134 136 138 140 142 144

Cents/kWh GWh

(29)

28 From the Italian cities only Palermo brings good results, with power output which exceeds the 100GWh per year in a low LCOE. In Cagliari and Brindisi, the main issue is the annual output which is very low. It is higher than Linear Fresnel mirrors, but still very low.

Figure 53. Base Case-LCOE for Italy Figure 54. Base Case- Annual Output for Italy

In this case the solar multiple does not refer to the power cycle, since it does not exist. The total number of the dish collectors will be changed. According to this number the solar field area and the electrical output of the solar field are changed automatically. Before the simulations it is important to present the solar multiple which is expressed in total number of collectors.

Table 2. Capacity and number of collectors expressed in Solar Multiple Solar Multiple Total Number of Collectors Nominal Electric Output(MW)

1 4000 100

2 8000 200

2.5 10000 250

3 12000 300

So the annual output and the LCOE with and without the influence of the incentives are shown below. The increase of the produced energy is very high as well as the GWh that Dish Stirling succeeds. It is quite interesting that the LCOE remain the same while the annual output increases sharply and this is the factor that makes this CSP system so competitive.

Figure 55. Solar Multiple variation-LCOE and Annual Output for Greece 59.2

31.6

54.5

0 10 20 30 40 50 60 70

59.3

111.6

64.4

0 20 40 60 80 100 120

31 31 31 31 35 35 35 35 40 43 43 43

23 23 23 23 26 26 26 26 31 31 31 31

0 50 100 150 200 250 300 350

400 LCOE (real-w/o

Cents/kWh GWh

Solar Multiple

(30)

29 The power output values exceed 400GWh and the LCOE are as low as in Power Tower. Especially in Spain, the climates have the potential to produce a great amount of electricity and these are the best results of all the CSP systems in this project. As it was mentioned before, the comparison is not valid since the solar multiples for 2, 2.5 and 3 do not refer in the same nominal power. But still the results for the base case are impressing.

Figure 56. Solar Multiple variation-LCOE and Annual Output for Spain

In Figure 57, the Italian simulations for different thermal storages are presented. Palermo has similar performance to Athens as it was noticed in the previous CSP systems. The power outputs and the LCOE without incentives are just a little bit higher in Athens. But the LCOE with incentive is quite the same. Cagliari has really low annual output in the case of 1 solar multiple. After the increase in the solar field, the output raises. The same happens to the Brindisi and both of them keep their LCOE stable like in the other locations.

Figure 57. Solar Multiple variation-LCOE and Annual Output for Italy 25 25 25 25 26 26 26 26 26 26 26 26 25 25 25 25

18 18 18 18 19 19 19 19 19 19 19 19 18 18 18 18 0

50 100 150 200 250 300 350 400

450 LCOE (real-w/o

59 59 59 59

32 32 32 32 54 54 54 54 43 43 43 43

23 23 23 23 40 40 40 40 0

50 100 150 200 250 300 350

400 LCOE (real-w/o

incentives) Annual Output (GWh) LCOE real Solar Multiple

Solar Multiple

(31)

30 Determination of the installed cost:

The installation cost in Dish Stirling for solar multiple of 1 is quite the same to this of the Linear Fresnel Mirrors in the base cases. But it is much lower from the one of the Parabolic Trough.

This CSP technology brings the best power outputs in the lowest LCOE and with a relatively low capital cost.

Figure 58. Installation cost for different values of Solar Multiple (S.M) 289.32

578.65 723.31

867.97

0 200 400 600 800 1000

1 2 2.5 3

Solar Multiple

Installation Cost

Million $

(32)

31

2.2.5 Planar PV Power Plants

The Flat plate PV system and its components are based in Sandia models for PV modules and inverters (SAM Help Contents NREL: Performance Models). Some physical properties of the modules are presented in Table 3.

Table 3. Module’s property characteristics Physical characteristics

Material Mono-c-Si

Efficiency 18.71%

Max Power Voltage 54.7V Max Power Current 5.58A

Max Power 305.226W

Module Area 1.631 m2 Number of Cells 96

It is also important to refer to the PV array sizing. Table 4 includes some interesting parameters about the desired layout.

Table 4. Parameters for the desired Layout Array Sizing

Total Number of Modules 327624 Modules per String 8 Strings in Parallel 40953 Number of Inverters 301 Desired Array Size 100MWdc

The next simulations are provided in order to be compared with the previous CSP systems. In the case of the photovoltaic PV the Annual Output decreases gradually during the lifetime of the project.

Since the useful life of the installation is 30 years, they will be presented three values of the Annual Output for each location. They will correspond to the 1st, the 15th and the 30th year. It is also significant to highlight that the Array Size is 100MW.

To begin with Greece, the LCOE are really low and the annual outputs relatively high, especially for Athens and Thessaloniki. Even after 30 years they exceed the 100GWh. In this system there are two points which should be noticed. Thessaloniki and Athens do not have a great difference in their annual outputs when in all the CSP systems the difference was significant and Athens’ results were far better than Thessaloniki’s. In Andravida the results are not as good as the other cities, but they are satisfactory and much better than in most of the CSP systems base cases.

Figure 59. LCOE with and without incentives for Greece

22.2 22.8

25.4

21.2 21.8

24.3

19 20 21 22 23 24 25 26

LCOE (real-w/o incentives) LCOE real Cents/kWh

(33)

32 Figure 60. Annual Output for the 1st, 15th and 30th year for Greece

In Spain the most appropriate location for a PV plant is Alicante and then Malaga. The third in the queue is Sevilla and finally Toledo. But the LCOE and the prices show that the performance in the four climates in good and the results are quite similar. Toledo is the last in performance due to its high elevation, but more details about this will be discussed in the Conclusion and discussion section. During 30 years lifetime, the PV lose more than 20GWh of their power output.

Figure 61. LCOE with and without incentives for Spain 148 138 128 144 134 125 129 121 112

0 20 40 60 80 100 120 140 160

Annual Output (GWh)

19.9

20.3

20.7

19.6 19.0

19.4

19.8

18.7

17.5 18 18.5 19 19.5 20 20.5 21

Years GWh

(34)

33 Figure 62. Annual Output for the 1st, 15th and 30th year for Spain

In Figure 63 and 64 are the results for the Italian cities. The PV system in the three Italian locations performs much better that it was expected. Palermo LCOE is better than these of the Greek cities and its power output is higher, even from Athens which always performed better in the CSP systems.

Cagliari and Brindisi did not perform well in any of the simulated CSP systems, but here their results are not so bad.

Figure 63. LCOE with and without incentives for Italy 166 154

143

162 151 140

159 148 137

168 157 145

0 20 40 60 80 100 120 140 160 180

Annual Output (GWh)

27.7

21.8

27.5 26.5

20.8

26.3

0 5 10 15 20 25 30

Years GWh

(35)

34 Figure 64. Annual Output for the 1st, 15th and 30th year for Italy

The Total Installed Cost for this plant is US $ 370.52 million.

119 111 103

151 151

130 120 112 104

0 20 40 60 80 100 120 140 160

Annual Output (GWh)

Years GWh

(36)

35

2.2.6 High-X Concentrating PV

An other photovoltaic PV system which will be examined is the High-X Concentrating PV. This type of system requires a higher capital cost due to the trackers and it also requires a cooling system in order to avoid losses. Despite of that, there are savings in land area and in PV modules. For the simulated installation the studied modules have the next characteristics which are presented in Table 5.

Table 5. High Concentration Photovoltaic Module’s property characteristics Physical characteristics

Maximum Power 375.553kWdc

Module Efficiency 29.8058%

Number of Cells 20

Concentration Ratio (lens area/cell area) 700X

Module Area 1.4m²

The array configuration is described in Table 6. Again it is important to mention that the system nameplate capacity is 100MWdc.

Table 6. Array Configuration Array Sizing Number of Trackers 1780 Modules on each Tracker 150 Number of Inverters 293 System Nameplate Capacity 100MWdc

The LCOE and the annual output are presented in Figure 65 and 66. To begin with the power outputs are lower than these of the PV plant. Also the LCOE is high enough and very close to some of the base cases of CSP systems, like the parabolic trough. The results of the outputs for Athens and Thessaloniki are not bad, but the further decrease of the produced energy leads to lower outputs, which does not make them competitive in long run.

Figure 65. LCOE with and without incentives for Greece

61.9 70.2

86.2

41.2 46.8 57.4

0 20 40 60 80 100

(37)

36 Figure 66. Annual Output for the 1st, 15th and 30th year for Greece

In Spain the LCOE are not so high and this time Sevilla has the worst performance. Also, the annual outputs are satisfactory. The PV plant reaches much higher values. It is noticeable that Alicante produces energy of 154GWh the first year of its operation and Malaga 149GWh, which are closer to the 146GWh of Toledo. So Malaga does not perform so well in the Concentrating PV system.

Figure 67. LCOE with and without incentives for Spain

Figure 68. Annual Output for the 1_st, 15_th and 30_th year for Spain

113 106 98 100 93 86 81 76 70

0 20 40 60 80 100 120

Annual Output (GWh)

47.1 52.0 47.9 45.6

31.4 34.7 31.9 30.4

0 10 20 30 40 50 60

LCOE (real-w/o incentives) LCOE real

149 139 129 135 126 116 146 136 127 154 143 133

0 50 100 150 200

Annual Output (GWh) Cents/kWh

Years

Years GWh

GWh

(38)

37 The last simulations are about Italy and they are presented in Figure 69 and 70. The results are disappointing for all of the locations. Even in Palermo which exceeded Greece outputs in the PV plants simulations. Actually, the power outputs are quite the same between Palermo and Athens.

The total installed cost is US $ 170.58 million higher than the PV plant and the results are much worse for all the countries and all the climates.

Figure 69. LCOE with and without incentives for Italy

Figure 70. Annual Output for the 1st, 15th and 30th year for Italy

The Total Installed Cost for this plant is US $ 541.1 million.

102.6

62.3

95.7 68.4

41.5 63.7

0 50 100 150

LCOE (real-w/o incentives) LCOE real

68 64 59

113 105 97

73 68 63

0 20 40 60 80 100 120

Annual Output (GWh) Cents/kWh

Years GWh

(39)

38

3. Results

In this section a detailed analysis of the simulations is presented. This includes matrices in the aid of which the results will be determined. The comparisons will be made according to the following sequence. The first step is to select the system with the best performance for every CSP technology.

Then for each location a certain CSP system will be optimized. The best CSP system variation for every climate will be compared both with the base cases and the other locations. After that, the optimized CSP systems for each location will be compared to determine the most suitable and cost effective system.

The procedure of the comparison will be made according to the following criteria. In the next matrices the best results for each variation are presented. The determination of the optimum solar multiple will be made according to the power output and LCOE. The case which reaches the maximum annual output and the lower LCOE will be the best one. In many cases the results are clear since the LCOE reaches its minimum for a certain value of the solar multiple.

In the case of the selection of the thermal storage capacity things are quite different, as almost all the values for the power output are proportional to the storage hours increase. The decision will be made according to the maximum difference between the LCOE and the annual output.

Moreover, they will be presented some more possible cases. In the Linear Fresnel Mirrors, there is the fossil fraction combination. This case is not presented in order to be compared with the other technologies but in order to be discussed and to show the fluctuation of the LCOE and the annual output.

In the Dish Stirling case, the base case corresponds to a nominal electric output of 100MW and can be compared with the other CSP technologies. Also, in the related matrices the case of the optimum solar multiple is presented as well. But it is not comparable. This is because an increase in the Solar Multiple is accompanied with an increase in the nominal electric output. From the related simulations, the solar multiple increase is proportional to the nominal electric output. It is important to mention that the LCOE (with incentives and without) remains constant with the solar multiple variations. So the best results are those of the higher solar multiple, which is 3.

Tables 19 and 20 show the base cases comparison and Tables 21 and 22 present the optimum scenarios comparison. It is significant to mention that the value of the Annual Output for the Flat Plate PV is this of the 15 year, which is supposed to present an average value of the Annual Output during the 30 years life time of the plant.

(40)

39

3.1 Parabolic Trough

Table 7. Comparison of Parabolic Trough’s simulations for Greece GREECE

Units Base Case Optimum S.M

Optimum T.S (S.M=1)

Optimum T.S (S.M=2.5) Athens

LCOE without incentives

Cents/kWh

65.42 (2.5) 38 (4h) 59 (12h) 43 LCOE with

incentives

Cents/kWh - - - 30

Annual Output

GWh 92.39 248 92 260

Thessaloniki LCOE without incentives

Cents/kWh 79.92

(2.5) 44 (4h) 72 (6h) 44 LCOE with

incentives

Cents/kWh - - - 30

Annual Output

GWh 75.56 214 76 214

Andravida

LCOE without incentives

Cents/kWh 90.69

(2.5) 51.8 (4h) 82 (12h) 59 LCOE with

incentives

Cents/kWh - - - 40

Annual Output

GWh 66.56 182 67 191

Installation Cost

Million $ ^514.53 ^833.17 (2.5)

456.63 (4h) 833.17 (6h) 1006.89(12h)

There are some points which do worth to be mentioned for the three Greek locations. The LCOE with incentives is the same for Athens and Thessaloniki in their optimum thermal storage hours for the same solar multiple (2.5).The LCOE without incentives is quite the same too. But the most important difference is in their annual output, which in Athens is 46 GWh more than Thessaloniki.

Also, their optimum case corresponds in different storage hours. It is noticeable that the great decrease in the LCOE comes after the increase of the solar multiple.

On the contrary, Andravida’s results are quite bad. The unit cost of electricity is much higher than in the other climates. Despite the modifications in the solar field and the storage hours, the LCOE does not decrease enough. More specifically, it takes the same value as in Athens’ system for a solar multiple of 1. But parabolic trough in Athens receives further improvement, which Andravida’s does not.

To sum up, the capital cost for an installation with 12 hours storage capacity is US $ 173.72 higher than that of 6 hours storage. This expensive investment needs to receive high power outputs and climate is the key factor for it. Andravida is not definitely the location which can come up with these expectations. The other two cities have more efficient parabolic trough systems.