Study of a hybrid concentrating solar power plant for Portuguese conditions
Full text
Study of a hybrid concentrating solar power plant for Portuguese conditions Bruno André da Costa Coelho Dissertation presented for the degree of: Doctor in Mechanical Engineering by the University of Porto Porto 2014 Armando Carlos Figueiredo Coelho de Oliveira, Adélio Miguel Magalhães Mendes, Joaquim Gabriel Magalhães Mendes, Study of a hybrid concentrating solar power plant for Portuguese conditions Bruno André da Costa Coelho Dissertation presented for the degree of: Doctor in Mechanical Engineering SUPERVISORS: Armando Carlos Figueiredo Coelho de Oliveira, Associate Adélio Miguel Magalhães Mendes, Associate Professor Joaquim Gabriel Magalhães Mendes, Auxiliary Professor Study of a hybrid concentrating solar power plant for SUPERVISORS: Associate Professor Associate Professor Auxiliary Professor
Acknowledgments I am grateful to the Portuguese Calouste Gulbenkian Foundation for their belief in my capabilities and financing my PhD grant (Grant Ref.: 104299); to UNET, FEUP and DLR for providing the conditions to develop my work. I will always be thankful to my supervisors Prof. Armando Oliveira, Prof. Adélio Mendes and Prof. Joaquim Gabriel for their guidance, for always being available (even when there was no time) and for our long discussions and brainstorming, which made possible to build this thesis. I thank my UNET labmates Szabolcs Varga and João Soares for their friendship; and to Peter Schwarzbözl and Klaus Hennecke for welcoming me in DLR. Also, thanks to all my friends for the coffees, the talks and the laughs that made all the difference. A special thank you to my father and my mother for always being there for me, for the unconditional help, for all the values and advices you gave me, which made me grow into the person I am today, and still do... Many thanks to my family for all moments we spent together, to my grandparents, my godmother Gustinha, my godson Alfredo, all my uncles, aunts and cousins. Last but definitely not least, thank you Eli for the love, patience and support during all these years, especially in the most difficult moments, gmdt.
In memory of my godfather Avelino
i. Contents Study of a hybrid concentrating solar power plant for Portuguese conditions Page ix Contents Abstract ................................................................................................................ xiii Sumário ................................................................................................................. xv Figure captions ...................................................................................................... xvii Table captions ...................................................................................................... xxiii Acronyms and abbreviations ................................................................................. xxv Codes ................................................................................................................... xxvi Currency exchange rates and Units ...................................................................... xxvii Variables ............................................................................................................. xxvii 1. Introduction ...................................................................................................... 3 1.1 Concentrated solar power (CSP) and Biomass outlook................................................. 4 1.2 Concentrated solar power (CSP) ................................................................................... 5 1.2.1 Why CRS? .............................................................................................................. 6 1.2.2 Central receiver system (CRS) technologies .......................................................... 7 1.2.3 CRS components.................................................................................................. 13 1.3 Why CRS hybridization? .............................................................................................. 27 1.3.1 Incentives in Portugal for CSP and biomass ........................................................ 28 1.3.2 SOLMASS Project ................................................................................................. 31 1.4 Biomass integration: resources and conversion technologies .................................... 32 1.5 Thesis structure, background and outputs ................................................................. 40 References ............................................................................................................. 41 2. Model of an atmospheric volumetric central receiver system (CRS) .................. 47 2.1 CRS solar field model ................................................................................................... 47 2.1.1 HFLCAL - Heliostat Field Layout Calculation ........................................................ 50 2.1.2 Annual simulation ............................................................................................... 54 2.1.3 Main factors that influence the solar field layout ............................................... 54 2.1.4 Distribution/optimization of the solar field ........................................................ 59 2.2 Power circuit model .................................................................................................... 71 2.2.1 Ebsilon Professional ............................................................................................ 71 2.2.2 Power circuit optimization approach .................................................................. 72 2.2.3 Steam cycle .......................................................................................................... 76 2.2.4 Air cycle ............................................................................................................... 84
ii. Sumário Study of a hybrid concentrating solar power plant for Portuguese conditions Page xvi “objetivo renovável", pode ser mantido. Se geridas de uma forma sustentável, as centrais híbridas podem também ter um impacto positivo na redução dos incêndios florestais e abrir novos mercados para o CSP. A integração de biomassa no circuito de vapor e de ar da CRS foi estudada no âmbito do projeto SOLMASS. A integração de biomassa no ciclo de vapor é mais favorável para a central híbrida de biomassa (FRB4#CRS#12). Esta solução híbrida reduz a energia desperdiçada e os períodos de arranque e paragem, reduzindo o LEC para 0,146 €/kWh. O LEC da central híbrida FRB4#CRS#12 é 0,086 €/kWh inferior à central solar CRS#12 e tem um aumento de 0,041 €/kWh quando comparado com uma central de biomassa (FRB4), reduzindo o consumo de biomassa em 7500 toneladas anuais. Considerando as tarifas nacionais atuais, a central híbrida FRB4#CRS#12 apresenta uma TIR de 6,6% em comparação com 9,7% de uma central solar CRS#12 e 7,4% da central de biomassa FRB4. A integração da biomassa no ciclo de ar de uma CRS pode ser feita considerando-se um amplo espectro de tecnologias: gaseificação de pellets ou resíduos de madeira, gaseificação de RDF, biogás gerado a partir de um digestor anaeróbico de águas residuais, biogás de aterro e gás natural. A solução com menor LEC foi obtida para a hibridização da CRS de 4 MWe com biogás de um digestor anaeróbico, utilizando lamas de uma estação de tratamento de águas residuais (LEC de 0,15 €/kWh). Esta central tem retorno do investimento em 13 anos (assumindo coleta e transporte de lamas sem custo) com o melhor VAL (15 milhões de euros) e TIR de todas as opções híbridas. Outra configuração interessante é a central híbrida CRS/gasificação de resíduos de madeira (WG#CRS#3), com um LEC de 0,17 €/kWh, tornando esta tecnologia de gasificação viável para Portugal, o que não acontecia no caso base. A energia solar concentrada (CSP) tem provado ser válida não só para a geração de eletricidade, mas para gerar combustíveis e produtos químicos a partir de fontes renováveis como a água, biomassa e sol. A revolução tecnológica na produção de hidrogénio e eletricidade é importante para apoiar as necessidades futuras e levar o mundo em direção a um futuro mais sustentável. Para isso, vários projetos em curso são apresentados e comparados. Embora falte ainda tomar medidas para resolver as limitações atuais e aumentar a viabilidade técnica e económica dos projetos, há condições para começar essa revolução e criar pontes das atuais tecnologias fósseis para as tecnologias renováveis.
iii. Figure captions Study of a hybrid concentrating solar power plant for Portuguese conditions Page xvii Figure captions Figure 1.1: Central receiver system, parabolic trough, dish/Stirling engine, linear Fresnel, Solar furnace. ......................................................................................................................................... 5 Figure 1.2: CRS pioneer projects: Eurelios, IEA-CRS e CESA-1; Solar One; Themis . ..................... 8 Figure 1.3: CRS power plant scheme. ............................................................................................ 9 Figure 1.4: PS10 e PS20 and power plant scheme. ....................................................................... 9 Figure 1.5: BrightSource CRS and power plant scheme. ............................................................. 10 Figure 1.6: Sierra SunTower and power plant scheme. .............................................................. 10 Figure 1.7: Solar tower of the Weizmann Institute, power plant scheme and the Masdar Institute concentrator. ................................................................................................................ 11 Figure 1.8: Gemasolar power plant and scheme. ....................................................................... 11 Figure 1.9: Jülich solar tower and power plant scheme ............................................................. 12 Figure 1.10: SOLGATE project volumetric pressurised air receiver and possible configuration for power plant integration .............................................................................................................. 12 Figure 1.11: Scheme of a particle CRS ......................................................................................... 13 Figure 1.12: Abengoa PS 10 - Sanlucar 120 heliostat and Esolar heliostat. ................................ 14 Figure 1.13: PS10 receiver and scheme. ..................................................................................... 15 Figure 1.14: Solar One receiver and scheme. ............................................................................. 16 Figure 1.15: Volumetric atmospheric receiver schemeand its cups. .......................................... 17 Figure 1.16: SOLGATE receiver and high temperature module. ................................................. 18 Figure 1.17: Salts and Andasol two-tanks storage solution. ....................................................... 20 Figure 1.18: Concrete storage systems. ...................................................................................... 21 Figure 1.19: Ceramic modules and integration into the Jülich CRS. ........................................... 22 Figure 1.20: Interior and exterior of a PCM storage module. ..................................................... 23 Figure 1.21: 4 steam tanks storage solution used at PS10 ......................................................... 23 Figure 1.22: Rankine cycle and its application to CRS ................................................................. 24
iii. Figure captions Study of a hybrid concentrating solar power plant for Portuguese conditions Page xviii Figure 1.23: Brayton cycle and application into CRS . ................................................................. 25 Figure 1.24: Jülich power plant air cooled condensers and hybrid condenser. .......................... 27 Figure 1.25: Biomass power plants and connection points approved under the 2006 Portuguese strategic programme. .............................................................................................. 30 Figure 1.26: SOLMASS solar-chemical concept. .......................................................................... 32 Figure 1.27: Main biomass conversion routes. ........................................................................... 34 Figure 1.28: Mortágua biomass power plant and Vila Velha de Ródão power plant ................ 36 Figure 1.29: Harboøre and Güssing power plants. ...................................................................... 37 Figure 1.30: Chianti power plant and scheme ............................................................................ 38 Figure 1.31: Barlavento landfill and Sotavento landfill ............................................................... 39 Figure 1.32: Vilamoura and Frielas WWTPs. ............................................................................... 39 Figure 2.1: Approaches for calculating the solar image incident in the solar receiver ............... 48 Figure 2.2: Image reflected by a heliostat with normal error distributions and real image reflected by the same heliostat .................................................................................................. 52 Figure 2.3: Image reflected by a spherical concentrator ............................................................ 53 Figure 2.4: Cosine factor and effect on the solar field ................................................................ 56 Figure 2.5: Blocking and shading losses and usual effect on the solar field. ............................. 57 Figure 2.6: Impact of atmospheric attenuation on the efficiency of a solar field for HFLCAL and measured data from Solar Two ................................................................................................... 58 Figure 2.7: Effect of atmospheric attenuation in the solar field spillage and the respective effect of the solar field. ............................................................................................................... 59 Figure 2.8: Heliostat scheme and dimension .............................................................................. 60 Figure 2.9: Typical heliostat layout for different algorithms: bilinear expanded, bilinear with spacing and slip planes. ............................................................................................................... 61 Figure 2.10: Spacing between heliostats in the solar field. ........................................................ 62 Figure 2.11: Spacing between heliostats to avoid blockings and examples for different distances to the solar tower. ....................................................................................................... 63 Figure 2.12: Starting values of ar in function of br for different heliostat sizes. ........................ 63
iii. Figure captions Study of a hybrid concentrating solar power plant for Portuguese conditions Page xix Figure 2.13: Non optimized solar field: highlighted are the best performance heliostats and their efficiency. ............................................................................................................................ 65 Figure 2.14: Tower height for different commercial CRS ............................................................ 66 Figure 2.15: Measured operating efficiencies of Jülich atmospheric air volumetric receiver. ... 67 Figure 2.16: Strategies for solar field focusing: central point and several points on a centre line ..................................................................................................................................................... 68 Figure 2.17: Solar field optimization strategy. ............................................................................ 70 Figure 2.18: Ebsilon structure. .................................................................................................... 71 Figure 2.19: Power circuit optimization strategy. ....................................................................... 73 Figure 2.20: Definition of the best power plant operational temperature range. ..................... 75 Figure 2.21: Ebsilon steam turbine model. ................................................................................. 76 Figure 2.22: SIEMENS steam turbine SST-110 Model ................................................................. 77 Figure 2.23: Different HRSGs: vertical, horizontal, single pressure with steam drum and Benson type ............................................................................................................................................. 78 Figure 2.24: Single pressure HRSG temperature and energy transfer profile. ........................... 79 Figure 2.25: Babcock & Wilcox HRSG output steam temperature influence in the power circuit efficiency ..................................................................................................................................... 80 Figure 2.26: Air cooled condenser Ebsilon model. ...................................................................... 81 Figure 2.27: Rankine cycle with superheating scheme and respective Temperature-Entropy diagram ....................................................................................................................................... 82 Figure 2.28: Rankine cycle re-heating scheme and respective Temperature-Entropy diagram . 83 Figure 2.29: Typical summer day history of available power from the solar field and generated electricity as a function of different solar multiples ................................................................... 85 Figure 2.30: Scenario for CSP potential and Portuguese electricity consumption - REN typical day ............................................................................................................................................... 86 Figure 2.31: Ebsilon model for the receiver and storage. ........................................................... 88 Figure 2.32: Ebsilon storage temperature profile. ...................................................................... 89 Figure 2.33: Ebsilon model for the receiver and storage. ........................................................... 90 Figure 2.34: Compilation of tools used for the power plant design and annual simulation. ..... 91
iii. Figure captions Study of a hybrid concentrating solar power plant for Portuguese conditions Page xx Figure 3.1: Cumulative hours per incident DNI for Faro, Portugal. ........................................... 103 Figure 3.2: Application of control strategy CS#1 to a typical operating day. ............................ 105 Figure 3.3: Decision diagram for CS#1 control strategy on a CRS. ............................................ 106 Figure 3.4: LEC variation with power block operating conditions ............................................ 109 Figure 3.5: LEC variation with receiver design DNI ................................................................... 111 Figure 3.6: Effect of solar multiple in the solar field performance for SM=1.25and SM=1.75 . 113 Figure 3.7: Influence of solar multiple, storage capacity and control strategy (CS#1 to CS#4) on CRS LEC ...................................................................................................................................... 114 Figure 3.8: Optimal power plant solar multiple (a) and optimal storage capacity (b) variations with control strategy ................................................................................................................. 115 Figure 3.9: Typical operational day for the 4 MWe CRS with 1.25 SM and 2 hours storage for CS#1 ........................................................................................................................................... 117 Figure 3.10: Performance of the 4 MWe CRS with 1.25 SM and 2 hours storage for CS#1. ..... 118 Figure 3.11: Cost structure of a 4 MWe atmospheric air volumetric CRS - option #3. ............. 119 Figure 3.12: Performance characteristics of the steam turbine. .............................................. 125 Figure 3.13: HRSG cost/area curves. ......................................................................................... 128 Figure 3.14: Possible impact of several factors in power plant LEC, comparing to reference case - CRS#3. ..................................................................................................................................... 129 Figure 3.15: Sensitivity impact of several factors in power plant LEC (compared to reference - CRS#3). ...................................................................................................................................... 130 Figure 3.16: Impact of possible mid-term innovations in LEC (compared to reference - CRS#3). ................................................................................................................................................... 131 Figure 4.1: Performance on a typical day for CRS#0. ................................................................ 142 Figure 4.2: Performance on a typical day for CRS#12. .............................................................. 143 Figure 4.3: Performance on a typical day for FRB4#CRS#0. ...................................................... 145 Figure 4.4: Performance on a typical day for FRB10#CRS#0. .................................................... 146 Figure 4.5: Performance on a typical day for FRB4#CRS#12_CS#5. .......................................... 147 Figure 4.6: Performance on a typical day for FRB4#CRS#12_CS#6. .......................................... 148
iii. Figure captions Study of a hybrid concentrating solar power plant for Portuguese conditions Page xxi Figure 4.7: Performance on a typical day for FRB4#CRS#12_CS#7. .......................................... 149 Figure 4.8: Performance on a typical day for FRB10#CRS#12. .................................................. 150 Figure 4.9: FRB10 Power plant configuration. .......................................................................... 153 Figure 5.1: Integration of biomass on CRS power plant air cycle, on a duct burner in the HRSG. ................................................................................................................................................... 163 Figure 5.2: Hybridization options for the considered CRSs. ...................................................... 164 Figure 5.3: WG power plant configuration – model for reference. .......................................... 179 Figure 5.4: RDF power plant configuration – model for reference. .......................................... 182 Figure 6.1: CSP routes for renewable hydrogen production. ................................................... 193 Figure 7.1: Schematic representation of REELCOOP prototype system 3 (hybrid CSP/biomass power plant). ............................................................................................................................. 217
iii. Figure captions Study of a hybrid concentrating solar power plant for Portuguese conditions Page xxii This page was intentionally left blank
iv. Table captions Study of a hybrid concentrating solar power plant for Portuguese conditions Page xxiii Table captions Table 1.1: Performance of several CSP technologies .................................................................... 6 Table 1.2: Heat rejection technologies and performance/cost impact ...................................... 26 Table 1.3: PIP results for CPV power plants ................................................................................ 28 Table 1.4: PIP results for CSP power plants ................................................................................ 29 Table 1.5: Biomass sources as fuel .............................................................................................. 33 Table 2.1: Software for optimization and distribution of heliostats on the solar field............... 49 Table 2.2: Characteristics of the mirrors available on the market. ............................................. 55 Table 2.3: Characteristics of the Helios3S heliostat. ................................................................... 61 Table 2.4: Starting values for heliostat distribution in the solar field. ........................................ 64 Table 2.5: HRSG evaporator pinch point and approach temperature ranges. ........................... 79 Table 2.6: Operating temperatures and flux ranges of CRS solar receivers. .............................. 84 Table 3.1: Design values for the power block cycle selection impact analysis. ........................ 102 Table 3.2: Power block operating conditions for a 4 MWe atmospheric air volumetric CRS. .. 108 Table 3.3: Design DNI and receiver flux impact in a 4 MWe atmospheric air volumetric CRS. 110 Table 3.4: CRS design conditions for the different solar multiples. .......................................... 112 Table 3.5: Cost distribution for power plants with different SMs. ........................................... 113 Table 3.6: Power plant cash flow analysis and economic indicators. ....................................... 116 Table 3.7: CRS component costs – option CRS#3 or CRS#3_SM1.25_2S_CS#1. ....................... 119 Table 3.8: ECOSTAR reference costs for an atmospheric volumetric CRS power plant. ........... 120 Table 3.9: Economic data for the CRS model. ........................................................................... 121 Table 3.10: Power block components cost - for best option. ................................................... 123 Table 3.11: Bare module factor for power block equipments. ................................................. 124 Table 3.12: Validation data for the SIEMENS SST-110 turbine. ................................................ 125 Table 3.13: Validation data for the three SIEMENS SST-060 turbines. ..................................... 127
iv. Table captions Study of a hybrid concentrating solar power plant for Portuguese conditions Page xxiv Table 3.14: Power block components cost - for best option. ................................................... 128 Table 3.15: Cash flow sensitivity analysis - CRS#3. ................................................................... 132 Table 4.1: Main routes considered for forest waste biomass burning plant integration on a volumetric open air CRS power plant at design point. ............................................................. 141 Table 4.2: Main results for biomass and CRS power plant steam integration. ......................... 144 Table 4.3: Economic analysis for hybrid power plants in the Portuguese Algarve region. ....... 151 Table 4.4: Validation of the biomass boiler model. .................................................................. 154 Table 5.1: Cost considerations for the base cases. ................................................................... 168 Table 5.2: Biomass and CRS base cases performance and cost. ............................................... 170 Table 5.3: Biogas and Syngas calculated compositions. ........................................................... 171 Table 5.4: CRS and biomass hybrid power plants performance and cost. ................................ 173 Table 5.5: Economic analysis for base cases. ............................................................................ 175 Table 5.6: Economic analysis for hybrid options. ...................................................................... 176 Table 5.7: Wood pellets and residues base case power plant and gasifier validation results. . 180 Table 5.8: Wood pellets and syngas generated - validation results. ........................................ 180 Table 5.9: Wood residues and syngas generated - validation results. ...................................... 181 Table 5.10: RDF base case power plant and gasifier validation results. ................................... 183 Table 6.1: Thermochemical “high temperature” water splitting cycles ................................... 200 Table 6.2: Thermochemical “low-temperature” water splitting cycles .................................... 201 Table 6.3: Hydrogen production cost and efficiencies per cycle. ............................................. 204
v. Acronyms and abbreviations Study of a hybrid concentrating solar power plant for Portuguese conditions Page xxv Acronyms and abbreviations BIPV – building integrated PV; CAPEX – capital expenditure; CHP – combined heat and power; CPC – compound parabolic concentrator; CRS – central receiver system; CS – control strategy; CSP – concentrated solar power; DC – direct costs; DE – dish/Stirling engine; DNI – direct normal irradiance; DLR – Deutsches Zentrum für Luft- und Raumfahrt (German Aerospace Centre) DP – Design point; DSC – dye-sensitized solar cells; EBSILON – Ebsilon Professional software; ECOSTAR – European concentrated solar thermal road-mapping; EPC – engineering, procurement and construction; ES – energy stored; EU – European Union; FEUP – University of Porto-Faculty of Engineering; FICFB – fast internal circulating fluidised bed; GREENIUS – the green energy system analysis tool; HFLCAL – Heliostat Field Layout CALculation; HP – high pressure; HRSG – heat recovery steam generator; HVAC – heating, ventilation, and air conditioning; HTF – heat transfer fluid; IC – indirect costs; IRR – internal rate of return; KAM – Kraftanlagen München GmbH; LEC – levelized electricity cost; LF– linear Fresnel; LP – low pressure;
1. Introduction Study of a hybrid concentrating solar power plant for Portuguese conditions Page 4 1.1 Concentrated solar power (CSP) and Biomass outlook Portugal has several important renewable resources: solar, geothermal, hydropower, biomass, wind, wave and tidal. Several studies estimate the renewable electricity generation based on the technical potential (which could be used for power generation in the present technological state of the art) and economic potential , for several countries in the region of southern Europe (including Portugal) and the Middle East / North Africa (MENA region) [1]. In the case of Portugal, the largest economic and technical potential is attributed to concentrated solar power (CSP). The CSP technical potential for Portugal is equal to 436 TWh/year and the economic potential to 142 TWh/year, which is sufficient to support twice the current national electricity consumption [1, 2]. The biomass economic potential is the second largest compared to all other technologies, with about 27 TWh/year. In terms of project implementation, the scenario changes in the case of Portugal. It is anticipated that the technology with the greatest economic implementation potential in 2050 has the lowest rate of execution, with about 7%, which would give a total electricity generated by CSP of about 10 TWh/year, about the same value that is expected to be produced based on biomass [1].
1. Introduction Study of a hybrid concentrating solar power plant for Portuguese conditions Page 5 1.2 Concentrated solar power (CSP) CSP is already described by the Greeks back to 213-221 BC as a mean for firing enemy wood made ships. However, CSP technologies for commercial deployment began only in the 1980s. Currently, there is a significant investment on CSP in the Mediterranean region, led by countries such as Spain, Algeria, Morocco, Israel and UAE, who are currently building large scale CSP power plants, with different technologies [3]. Also, in the USA, mainly in California and Nevada, several power plants have been recently inaugurated. There are four main CSP systems for electricity generation: central receiver systems, parabolic trough, dish/Stirling engine, and linear Fresnel. Although with minor expression, other solar concentrating systems can also be found, such as the solar furnace (used for chemical applications). These systems are illustrated in Figure 1.1. Figure 1.1: Central receiver system (CRS – up left), parabolic trough (PT – up centre), dish/Stirling engine (DE – up right), linear Fresnel (LF – down left), Solar furnace (SF – down right).
1. Introduction Study of a hybrid concentrating solar power plant for Portuguese conditions Page 6 The parabolic trough, linear Fresnel and central receiver systems can be coupled to steam cycles up to 200 MW of electrical capacity, with thermal cycle efficiencies around 30- 40%. On the other hand, the dish/Stirling engines are typically used for decentralised electricity generation, in the range 10 - 20 kW [1]. 1.2.1 Why CRS? From all the commercial solar concentration technologies the most mature is the parabolic trough. Despite that, in recent years, the technology with more R&D attention has been the central receiver technology, mainly because it can reach very high temperatures (over 1000 °C). New prototypes are capable of producing hot air for the gas turbine operation, which can be used in combined cycles, generating higher conversion efficiency, and approximating these power plants to commercial natural gas power plants, with power efficiencies above 50%, as shown in Table 1.1 [1]. Table 1.1: Performance of several CSP technologies [1]. Technology Installed Power (MW) Solar concentration Max. Sunelectricity efficiency Annual efficiency Power block efficiency Capacity factor Land Use (m 2 MWh -1 year -1 ) CRS 10 - 150 300 - 1000 20 % (d) 35 % (p) 8-10 % (d) 15-25 % (p) 30-40 % (ST) 45-55 % (CC) 25- 90 % (p) 8 - 12 PT 10 - 200 70 - 80 21 % (d) 10-15 % (d) 17-18 % (p) 30-40 % (ST) 24 % (d) 25-90 % (p) 6 - 8 DE 0.01 - 0.4 1000 - 3000 29 % (d) 16-18 % (d) 18-23 % (p) 30- 40 % (EN) 20-30 % (GT) 25 % (p) 8 - 12 LF 10 - 200 25 - 100 20 % (p) 9-11 % (p) 30-40 % (ST) 25- 90 % (p) 4 - 6 (d) = demonstrated, (p) = projected, ST – steam turbine, GT – gas turbine, CC – combined cycle, EN – Stirling engine. Solar efficiency = net electricity generated / solar field incident irradiation. Capacity factor = Solar hours or power plant operating hours / 8 760 hours per year. The most promising present solution in terms of efficiency is the CRS. Also, the solar energy can be concentrated into a selected focal point, which opens the application of CSP for a great number of applications, e.g. electricity or chemicals generation.
1. Introduction Study of a hybrid concentrating solar power plant for Portuguese conditions Page 7 1.2.2 Central receiver system (CRS) technologies CRS technology began its pioneering solar concentration existence with various projects between 0.5-10 MW in the early 1980s. Among these projects are: • The Eurelios solar plant (1980-1984) that was built by an Italian/French/German consortium funded by the Commission of the European Communities, in Adrano, Sicily. With an installed power of 750 kWe and molten salts storage, it used a water/steam cycle, reaching peak temperatures of 512 °C. The power plant had an estimated cost of 8.2 million U.S. dollars (Figure 1.2) [4, 5 and 6]; • The IEA-CRS (1981-1985) and CESA-1 (1983-1984) solar power plants were built with the support of the European Community in Almeria, Spain. The IEA-CRS power plant had an installed capacity of 500 kWe and used sodium for storage and operating fluid, reaching temperatures of 560 °C. The power plant had an estimated cost of 18 million U.S. dollars and worked with a peak solar to electricity efficiency of 8.1% (Figure 1.2) [4, 5 and 6]; • The Sunshine solar power plant (1981-1984) was built by the Japanese government, in Nio, Japan. It had an installed capacity of 800 kWe and used steam as working fluid and storage medium, reaching temperatures of 249 °C. The power plant had an estimated cost of 25 million U.S. dollars and worked with a solar to electricity efficiency of 9.2% [4, 5 and 6]; • The Solar One (1981-1986) and Solar Two (1995-1999) solar power plants were built by the North American Department of Energy, in the Mojave Desert, California. Solar One had an installed power of 11.7 MWe and used oil at 302 °C as working fluid and storage, operating the power block with steam and reaching temperatures of 510 °C. The power plant had an estimated cost of 141 million U.S. dollars and worked with annual average solar to electricity efficiency of 5.8% and a peak efficiency of 8.7%. This plant was demolished in 2009 (Figure 1.2) [4, 5 and 6]; • The Themis solar power plant (1983-1986) was built by EDF, in the region of Cerdanya, in the Pyrénées-Orientales, France. It has an installed capacity of 2.3 MWe and uses molten salts as working fluid and storage (in two tanks). The power plant had an estimated cost of 37 million U.S. dollars and worked with annual average solar to electricity efficiency of 17% (Figure 1.2) [4, 5 and 6];
1. Introduction Study of a hybrid concentrating solar power plant for Portuguese conditions Page 8 • The C3C-5 solar power plant (1985 -1988) was built by the Soviet Union government, in Crimea, now part of Ukraine. It had an installed power of 5 MWe and used steam as operating fluid and storage, reaching temperatures of 256 °C [4, 5 and 6]. Figure 1.2: CRS pioneer projects: Eurelios (up left), IEA-CRS e CESA-1 (up right); Solar One (up left); Themis (up right) [9, 10 and 11]. All of these different solar towers or central receiver systems may be described in terms of the following components: • Solar field, consisting of a number of mirrors with two axis solar tracking and optimally distributed through the field - heliostats; • Solar receiver, where the flow of concentrated solar radiation is absorbed; • The heat transfer system, where a heat transfer fluid (HTF) is used to carry thermal energy from the receiver to the turbine circuit; • Thermal energy storage (TES) system, which ensures the dispatchability of the system during periods of low radiation and allows adapting power to the demand curves; • Backup of fossil fuels/renewable resources for hybrid systems; • Power block, e.g. steam generator, turbine and electric generator;
1. Introduction Study of a hybrid concentrating solar power plant for Portuguese conditions Page 9 Figure 1.3 represents a simplified schematic for a solar tower/CRS solar concentration power plant. Figure 1.3: CRS power plant scheme. Many of the pioneering central receiver systems (Figure 1.2) no longer operate, and are now disabled or have been demolished. During their lifecycle many setbacks were registered in key components, such as heliostats and solar receiver. The subsequent improvement of all essential components, in many international projects during the last 30 years, resulted in an significant growth of CRS power plants in recent years and a diversification of the technologies, namely with the use of different heat transfer fluids. Today, the CRS commercial projects with more success, are the “Plantas Solares” PS10 (2007) and PS20 (2009), built by the Abengoa group in Sanlúcar la Mayor, Seville. The PS10 has an installed power of 11 MWe using saturated steam as heat transfer fluid and steam pressure storage (1 hour range), reaching temperatures of 250-300 °C. The estimated construction cost was 45 million U.S. dollars (Figure 1.4) [7, 8]; Figure 1.4: PS10 e PS20 (left) [7] and power plant scheme (right).
1. Introduction Study of a hybrid concentrating solar power plant for Portuguese conditions Page 10 Following this project, several companies have announced power plants of large dimension and using steam at much higher temperatures, such as Ivanpah Solar Electric Generating Station (2013), built by the company BrightSource Energy, Primm, California. BrightSource planned to build several plants in California, totalling 370 MWe, with superheated steam as heat transfer fluid, reaching temperatures around 565 °C, nearly twice the steam temperature used by Abengoa's PS10 and PS20 (saturated steam). The company recently acquired loan guarantees from the USA Department of Energy estimated in 1375 million U.S. dollars (Figure 1.5) [9]; Figure 1.5: BrightSource CRS (left) [9] and power plant scheme (right). A different approach was used by the company e-solar, also using saturated steam technology, in the Suntower Sierra (2009) project, at Lancaster – California. The 5 MWe power plant reaches temperatures of 440 ° C using a multi-tower/receiver concept with heliostats of small dimensions. The company hails costs competitive with conventional fossil fuel technologies, but did not disclose the final costs associated with this project. The same company has announced two more projects (in progress): Alpine Suntower (2012, 92 MWe) and New Mexico Suntower (to be determined, 92 MWe) (Figure 1.6) [8, 10]; Figure 1.6: Sierra SunTower (left) [8] and power plant scheme (right).
1. Introduction Study of a hybrid concentrating solar power plant for Portuguese conditions Page 11 An alternative CRS design is to use a secondary concentrator which concentrates the solar image into a receiver placed at ground level. These systems are called "beam down" and the two reference prototypes are the solar tower at the Weizmann Institute in Israel - 3 MWth and the 100 kW concentrator in the city of Masdar at Abu Dhabi, UAE (Figure 1.7) [11, 12]; Figure 1.7: Solar tower of the Weizmann Institute (left), power plant scheme (centre) and the Masdar Institute concentrator (right). A different heat transfer fluid used in power CRS are molten salts. The solar plant Gemasolar or solar three (2011), built by the group Sener in Fuentes de Andalucía (Seville) has an installed power of 17 MWe, two molten salt storage tanks (15 hour equivalent), and the heat transfer fluid reaches temperatures of 565 °C. This plant was built based on the experience gained from the design and operation of solar 1 and solar 2 power plants, in California. The Spanish group Sener announced an estimated funding of 309 million U.S. dollars for the construction and commissioning of the plant (Figure 1.8) [13]; Figure 1.8: Gemasolar power plant (left) [13] and scheme (right). The company United Technologies Corp. - Pratt Whitney Power Systems, recently inaugurated two large projects using a molten salt as heat transfer fluid: the project Crescent Dunes Solar Energy Project (2013, 100 MWe), built for Tonopah Solar Energy, LLC, in Tonopah,
1. Introduction Study of a hybrid concentrating solar power plant for Portuguese conditions Page 12 California; and the Rice Solar Energy Project (2013, 150 MWe), built for the Rice Solar Energy in Rice, California [10, 14]. The current largest commercial power plants use steam or molten salts as heat transfer fluids. A different concept is to use air (atmospheric and pressurized) as heat transfer fluid. Various prototypes and pre-commercial plants were already built. The atmospheric air technology current reference is the Jülich solar tower (2009), designed by DLR and the company KAM in Jülich, Germany. It has an installed capacity of 1.5 MWe, with a ceramic storage and atmospheric air as working fluid, reaching temperatures of 700 °C. The estimated cost of construction was approximately 32 million U.S. dollars (Figure 1.9) [15]; Figure 1.9: Jülich solar tower (left) [15] and power plant scheme (right). An alternative concept to the use of air as heat transfer fluid, is by "closing" the volumetric receiver, pressurizing the air and reaching higher temperatures, which enable the use of a more efficient combined cycle (Brayton-Rankine). The project with the highest relevance in this technology was the project SOLGATE, held in 2001 at the Plataforma Solar Almeria, which built and tested a full hybrid system with an installed power of about 250 kWe, Figure 1.10 [11]. Figure 1.10: SOLGATE project volumetric pressurised air receiver (left) [16] and possible configuration for power plant integration (right).
1. Introduction Study of a hybrid concentrating solar power plant for Portuguese conditions Page 13 There are still many other concepts and technologies for a CRS. One of the concepts with high interest for investigation and development is the use of particles (e.g., graphite) to absorb solar radiation and use downstream heat exchangers to feed a thermal cycle (e.g. with air - Brayton cycle). The particles are also used as high temperature heat storage medium. The air under pressure can reach temperatures of about 995 °C [17]. This concept is still in the research stage and there are no commercial or pre-commercial power plants using this technology, Figure 1.11. Figure 1.11: Scheme of a particle CRS [17]. 1.2.3 CRS components 1.2.3.1 Solar field The CRS solar field is composed by a set of concave mirrors, which follow the sun trajectory in two axes during the day, focusing and concentrating its direct normal irradiance (DNI) at a position set by the operator. The positioning and distribution of the heliostats in the solar field is dependent on the power plant location and the type of receiver used. There are three types of solar fields: only north of the solar tower (typically in positions in the northern hemisphere); only south of the solar tower (typically in positions in the southern hemisphere); and surrounding solar fields (typically in equatorial positions). The field type also depends on the size of the plant.
1. Introduction Study of a hybrid concentrating solar power plant for Portuguese conditions Page 20 • Storage using molten salts and ionic liquids; • Storage using concrete modules; • Storage using ceramics modules; • Phase change materials (PCM) storage; • Steam storage. In the following sub-chapters these technologies are presented and discussed. Storage using molten salts and ionic liquids The state of the art of molten salt storage technology is two-tank storage. This solution was extensively tested in Solar Two power plant, in combination with a molten salt solar tubular receiver. There are different salts under study, but the combination that best fits the temperature range of operation of a Rankine cycle is a mixture of 60% sodium nitrate and 40% nitrate potassium. This was the chosen solution for the Gemasolar (Solar 3) solar plant. The molten salt two-tank storage [25] was applied in the Gemasolar power plant, Figure 1.17, but was also used in parabolic trough systems as indirect storage. Andasol uses a heat exchanger to transfer heat from the oil to the salts. If molten salt were used as heat transfer medium, it may cause problems in the parabolic trough receiver tubes, because salts solidify at temperatures of 220 °C. Using oil directly as storage material might also be an alternative, however is not normally used because the vapour pressure of 12 bar and the high price. Figure 1.17: Salts (left) and Andasol two-tank storage solution (right) [25]. The use of ionic liquids can bypass this drawback, since these materials are liquid even at low temperatures. Ionic liquids are organic salts with negligible vapour pressure at the relevant temperature range and melting temperature below 25 °C. The ionic liquids at ambient temperature are still materials with unknown results to CSP, and it is uncertain that they are
1. Introduction Study of a hybrid concentrating solar power plant for Portuguese conditions Page 21 stable at the temperature level required for CSP systems, and also if they can be produced at a competitive cost [20]. Storage using concrete modules The concept of storing sensible heat in concrete has been studied by Wespe and Wanda projects for parabolic trough systems. The projects studied two possible solutions: concrete with inner tubes and tubeless. Since the use of steel pipes within the storage material is very expensive (accounts for 45-55% of total storage cost), storage without tubes may result in reduced costs, however still requires extensive research [20]. Advanced loading/unloading modes can significantly increase storage capacity for a given size and material. The basic idea of modular storage is to increase the storage capacity by changing between the two modes of operation. Extensive simulations have shown that in those designs the ability of a given storage size can be increased by about 200% over the base case operation [26]. Storage in concrete is highly modular and is easy to apply loading and unloading, creating different temperature modules. However, the implementation of a concrete storage system is still risky for both cases (with or without tubing) and require further investigation before commercial application, Figure 1.18. Figure 1.18: Concrete storage systems. Storage using ceramics modules The storage of sensible heat using solid materials (e.g. ceramics) is usually used in combination with atmospheric or pressurized volumetric receiver systems. In these high temperature systems, heat needs to be transferred to another carrier, which may be several types of solids, provided they have high heat capacity and density. The materials size and shape of the solids is also important, as an optimized size and shape would minimize pressure drop and increase heat transfer, thereby reducing energy consumption. Apart from solid storage materials, there are other emerging concepts such as the use of silica sand. The Jülich
1. Introduction Study of a hybrid concentrating solar power plant for Portuguese conditions Page 22 power plant has a 10 MWth ceramics storage, which provides storage heat at 700 °C for 1 equivalent hour of the power plant consumption. The layout of these ceramic materials and their integration in the solar plant are shown in Figure 1.19 [33, 34]. Figure 1.19: Ceramic modules (left) and integration into the Jülich CRS (right) [27, 28]. Another possibility of storage development for pressurized volumetric receivers are insulated storage containers resistant to pressure up to 16-20 bar (depending on the pressure ratio of the gas turbine) [20]. PCM storage Phase change materials (PCM) can be the future for heat storage, which is particularly important in direct steam generation systems, since the temperature remains constant in the enthalpy range of phase change. PCM storage does not use exclusively solid-liquid transitions, but may also use solid-gas or liquid-gas transitions. However, solid-liquid transition is the most studied and applied. The two main R&D areas of PCM storage are [29]: • Encapsulation of small amounts of PCM; • Incorporation of PCM into a matrix made of another solid material with high heat conduction. The first line of research aims to find a way to reduce the distances within the PCM, while the second line of research aims to increase the PCM heat conduction. The solid PCM is the limiting factor for heat transfer between the fluid and storage, once it agglomerates between the pipe and the PCM liquid phase. Developments in these areas will enable PCMs to overcome the early stages of development and move to the first prototypes in real-scale, after some initial laboratory experiments, Figure 1.20.
1. Introduction Study of a hybrid concentrating solar power plant for Portuguese conditions Page 23 Figure 1.20: Interior (left) and exterior of a PCM storage module (right). Steam storage The steam drum, which is a common part in many steam generators, is a type of storage as it contains a quantity of boiling water under pressure. Steam could be produced from that component only by reducing the pressure. This type of storage is used many times as heat storage process for several industries. The main problem is scaling to larger capacities and the degradation of steam quality during unloading. This type of storage is ideal for short periods of time, in the range of several minutes, to compensate for the solar area shading in the case of rapid passage of small clouds. Such a system is used in PS10 in Seville, with a thermal capacity of 20 MWh of saturated steam (equivalent to 50 minutes of power block operation at 50 % load). The system is composed by four tanks that are operated in sequence as they are loaded. During operation of the plant at full load, and according to the strategy defined by the operator, the steam at 250 °C and 40 bar pressure, from the receiver, is used to charge the thermal storage system. When energy is needed, it is recovered from the storage tanks (Figure 1.21) to feed the turbine [7]. Figure 1.21: 4 steam tanks storage solution used at PS10 [7]. In the future, the use of encapsulated PCMs within the storage tanks, could improve the storage capacity, because the latent heat content can be used to reduce the falling of temperature and pressure, thus allowing larger heat capacity with the same storage tanks [20].
1. Introduction Study of a hybrid concentrating solar power plant for Portuguese conditions Page 24 1.2.3.4 Power block Thermal cycles State-of-the-art CSP plant thermal cycles are based on the Rankine cycle, Figure 1.22 [30]; the water is pumped from a low pressure to a high pressure, with external power consumption. The pressurized water then enters a boiler/receiver (heat source) where it is heated at constant pressure until it becomes saturated/superheated steam. The saturated/superheated steam then expands through a turbine to generate work. With this expansion, both pressure and temperature are reduced. Finally, the steam enters a condenser where it is cooled to liquid condition. This water is then pumped and the cycle is repeated [30]. For a CRS, in the case of tubular steam receivers, the operation diagram is similar to Figure 1.22, because the receiver acts as a solar boiler and directly feeds the turbine. In the case of a molten salt tubular receiver, there is an intermediate step, the molten salt/ watersteam heat exchanger. This intermediate step is also applied in the volumetric receiver power plants that use a steam generator to recover heat from the hot atmospheric air and generate superheated steam to be fed to the turbine. Both molten salts and air technology have a decrease in efficiency due to this intermediate step, which is counterbalanced by better control. Figure 1.22: Rankine cycle (left) and its application to CRS (right) [30]. Commercially, the main power block companies have been developing series of turbines/generators specifically oriented towards CSP applications. This is the case of Siemens [31], Man Turbo [32] (Project Andasol 3) and General Electric [33] (Project PS10). In the case of pressurized volumetric receivers it is possible to operate a gas turbine under the Brayton cycle, Figure 1.23 [30], reducing the impact of the very inefficient compression of hot gases. Currently, this solution was only tested in prototype scale, and a
1. Introduction Study of a hybrid concentrating solar power plant for Portuguese conditions Page 25 large part of the power was obtained using a backup fossil fuel without recovering the gas turbine exhaust gases. Nevertheless, some pre-commercial power plants for this technology were already announced, as the case of the Themis power plant renovation [34]. Figure 1.23: Brayton cycle (left) and application into CRS (right) [30]. The main advantage of the Brayton cycle is that the remaining energy is rejected from the combustion in the form of heat in the hot exhaust gases, but can be recovered by a HRSG to run a Rankine cycle in what is called a combined cycle. These combined cycles could be the future of CRS technology, increasing efficiency of the power plant. However, turbine resistance to high temperatures is an extremely critical point for the technology viability, as well as turbine poor adaptability to transient heat flow situations. These issues deserve further R&D towards a full implementation of combined cycles in CRS power generation. Heat rejection The CRS technology usually consider two heat rejection systems: water and air cooled condensers. The use of water condensers leads to higher system efficiency, but also large water consumption (a problem in most countries with good CSP resources). The use of aircooled condensers usually requires an increase in electricity consumption, thus reducing the overall efficiency. It is also possible to use a hybrid scheme, with water and air cooling, but usually with an increase in the project cost, Table 1.22 [35].
1. Introduction Study of a hybrid concentrating solar power plant for Portuguese conditions Page 26 Table 1.2: Heat rejection technologies and performance/cost impact [35]. Technology Cooling Capacity (dm 3 /MWh) Performance decrease* Cost increase ** Coal / Nuclear Once through 87000 to 102000 Recirculation 1700 to 2800 Air 200 to 250 Natural gas Recirculation 750 CRS Recirculation 1900 to 2800 Hybrid 350 to 950 1 to 3 % 5 % Air 350 1.3 % PT Recirculation 3000 Hybrid 400 to 1750 1 to 4 % 8 % Air 300 4.5 to 5 % 2 to 9 % DE Washing 75 LF Recirculation 3800 * Loss in the annual electricity generation in comparison with the most efficiency technology; ** Additional cost to the electricity generation. Water cooling condensers were extensively used in the SEGS U.S. system, but consumed approximately 800-1000 litres of water per MWh generated; the Rankine cycle heat rejection system represents approximately 90 % of water consumption of the plant; and the other 10% of water consumption includes the water replacement in the steam cycle (8%) and water used for washing the solar field mirrors (2%) [36]. Commercial CRS power plants (Abengoa’s PS10 and PS20, Sener’s Gemasolar and e-solar’s Sierra Suntower) all use water cooling condensers, Figure 1.24. The use of air cooled condensers can only be found on the Jülich Solar power plant and on BrightSource Ivanpah power plant [37]. The Crescent Dunes Solar Energy power plant will be the first CRS commercial project using a hybrid heat rejection mechanism [37].
1. Introduction Study of a hybrid concentrating solar power plant for Portuguese conditions Page 27 Figure 1.24: Jülich power plant air cooled condensers (left) and hybrid condenser (right). 1.3 Why CRS hybridization? Central receiver solar only power plants have some limitations, inherent to the solar irradiance characteristics: transient daily operation; unavailability of energy resource during the night; high capital investment per installed power; and difficulty in market entry. This restrictions lead to low capacity factors or over-dimensioned and expensive storage devices. Overall, and because of these issues, CSP and CRS have quite high levelized electricity costs. Hybridization can be a solution to these concentrating solar power shortcomings, providing energy for night operation or to support transient daily operation, increasing the power plant capacity factors and reducing market entry difficulties. CRS hybridization can be done with different objectives: for electricity generation, for chemicals or fuels generation or even for fresh water generation. Several resources can be used for this hybridization, e.g. fossil fuels or renewable resources such as: biomass, geothermal, photovoltaic or wind. Each solution has a specific approach and current different R&D status. Hybridization with geothermal energy is possible but more restrict, since good CSP coincident with good geothermal spots around the world are limited, and are usually subjected to frequent earthquakes that are not good to CRS, because it affects the heliostat structure and misaligns the reflected irradiance. Fossil fuels and biomass are very interesting hybridization possibilities, as the electricity conversion cycles are similar to the CRS ones, and they can answer to the major limitations from CSP. Both solutions are studied in this thesis, with the main emphasis in biomass hybridization, so the “renewable goal” is maintained. The hybridization makes sense also from the biomass perspective. As it will be presented in the following chapters, biomass only power plants shortcomings can be minimized in hybrid CRS/ biomass power plants.
1. Introduction Study of a hybrid concentrating solar power plant for Portuguese conditions Page 28 1.3.1 Incentives in Portugal for CSP and biomass There are no incentives for hybrid CSP biomass power plants in Portugal. The network integration (PIP) calls are specific either for CSP or biomass power plants. Nevertheless, the Portuguese feed-in tariff for renewable energy projects is calculated by a formula which can consider contributions from different resources. Regarding CSP, a call was opened in Portugal for network integration (PIP) of CSP and concentration photovoltaic (CPV) power plants on September 2009, by the order no. 18838/2009, published on the national republic journal DR n.157, 2 nd Series, from the 14 th August 2009 [38]. 87 proposals were presented and 65 were accepted, but only 15 were selected. From these projects, 5 CPV projects were selected totalling 5 MW of installed capacity, Table 1. [38]. Table 1.3: PIP results for CPV power plants [38]. Promoter Location Power (MWe) Technology Reciclamas, SA Tavira 1 SOLFOCUS SAPEC – Química, SA Sapec 1 MAGPOWER Tecneira – Tecnologias Energéticas, SA Alqueva 1 OPEL LUZ.ON – Solar Energy, SA LUZ.ON 1 CONCENTRIX + AMONIX Glintt – Global Intelligent Technologies Évora 1 EMCORE The intention of national authorities with the PIP was to create a commercial demonstration platform for the main CSP technologies available, approving 4 projects with dish/Stirling engines, 2 projects with the parabolic trough technology, 2 projects with linear Fresnel technology and 2 projects with central receiver technology. The total installed power approved for dish/Stirling engines was 5.5 MWe while the other 6 projects total 24 MWe; divided into 6 power plants of 4 MWe installed power each, Table 1.4 [39].
1. Introduction Study of a hybrid concentrating solar power plant for Portuguese conditions Page 29 Table 1.4: PIP results for CSP power plants [38]. Promoter Location Power (MWe) Technology Ramada Holdings, SGPS Évora 1.5 DE Hyperion Energy Portugal Évora ou Reguengos 1.5 DE Selfenergy Silves 1 DE Bragalux Évora 1.5 DE Efacec Tavira 4 CRS Abengoa/Fomentinvest Moura 4 CRS Energena SLU Évora 4 PT Martifer Energia Évora 4 PT Dalkia Faro 4 LF Tom Moura 4 LF The Faculty of Engineering, University of Porto (FEUP) was the national scientific partner for the project led by EFACEC to build a 4 MWe CRS in Tavira. The CRS technology uses an atmospheric air volumetric receiver, and included the participation of the German Aerospace Center (DLR) as international scientific partner and the company Kraftanlagen München (KAM) as engineering, procurement and construction (EPC) contractor. FEUP gathered the consortium that applied to the call and was responsible for the call’s technical proposal. The EFACEC-FEUP proposal got the first position in the CRS call, and this thesis work was built to support the construction of a 4 MWe atmospheric air volumetric CRS power plant in Portugal, developing innovative models for the SOLMASS project and proposing innovative options for the hybridization with biomass. For the location selected, CRS and biomass hybridization can also interesting from the government perspective. It could allow reaching the objectives set for the renewable electricity strategic plan. In 2006, Portuguese authorities launched a PIP call to support biomass power plants grid connection allocating up to 100 MWe (equivalent to the annual consumption of one million tons of biomass), mainly in areas of with large biomass resources and risk of fire [40]. Following the call, connection points were defined and several projects were approved, Figure 1.25 [41].
1. Introduction Study of a hybrid concentrating solar power plant for Portuguese conditions Page 36 plant, with a capacity of 4 MWe [51] (operating with forest residues, bark residues, pine sawdust and olive bagasse), Figure 1.28. Figure 1.28: Mortágua biomass power plant (left) and Vila Velha de Ródão power plant (right). The use of forest biomass for energy production could have negative and positive effects. On one hand, the energy agriculture may be an excellent opportunity to promote sustainable agriculture and low ecological impact, and could be an incentive for the forest products industry to manage their resources more efficiently and thus improve the health of the forest. But it can also provide an excuse under the "green" cover, to explore the forests unsustainably, as unfortunately happened in the past, and many people regard with alarm the prospect of increased wood logging. The solution is to analyse the use of biomass integrated into forestall policies, so that integration in the range of renewable energy is successful. If the source is agricultural, the use of biomass should also be carefully examined to avoid collisions with food industry, especially for surface area required and the selected sources. A massive and inadequate use of biomass for energy purposes can compete with agriculture and possibly increase the basic food prices or influence the quality of soil and agricultural economies. However, if done in a sustainable manner, energy crops can provide a steady income to supplement farmers in periods between stations, stabilize crops susceptible to erosion and flooding, or allow farmers to work without requiring much unproductive additional equipment. Thus, there is also a need for regulation and supervision of the development of energy crops and waste. With regard to emissions of greenhouse gases, current forest biomass plants generally have similar emissions as coal plants, with the notable difference that biomass emits very little sulphur dioxide and toxic metals (cadmium, mercury and others). The most serious problem is the emission of particulate matter that must be controlled with special equipment. One of the greater environmental benefits of replacing fossil fuels with biomass is that, if done in a sustainable manner, the amount of carbon dioxide emitted when biomass is burned is almost the same as it is absorbed by the biomass growing up, thus forming a sustainable fuel cycle
1. Introduction Study of a hybrid concentrating solar power plant for Portuguese conditions Page 37 with virtual none global emissions of carbon dioxide. Some entries of fossil fuel may be needed for growing, harvesting, transportation and processing of biomass. However, the emissions from cultivation should be small, and if the energy to produce and process biomass comes from renewable sources, the net contribution to global warming would be nearly zero. The most advanced biomass technologies such as gasification should generate very low emissions, comparable to natural gas plants. Biomass gasification processes have a wide range of applications and higher energy conversion efficiency [52, 53] than biomass combustion power plants. Biomass gasification can produce a mixture of methanol, heat, synthesis gas and electricity with efficiencies up to 73% [53]. The wood biomass, although more costly than the forestry residues, has better and more consistent properties which are beneficial to feed a gasifier. Under these conditions most of the problems associated with the fluidized bed gasifier or tars formation are relatively well known, so the biomass gasification industry already has commercial solutions to these issues. There are however few commercial biomass gasification to electricity generation power plants. Examples of demonstration and pre-commercial projects can be found in the Nordic countries and Central Europe. The 3.5 MW gasification of wood shavings Harboøre power plant is in operation since 1996 and has over 8000 hours of annual operation [54]. In the power plant early days, some problems emerged mainly in the cleaning of the synthesis gas. However, the company that built the process in 2000, Babcock & Wilcox Vølund, solved the problems and at the end of 2003 the general concept of biomass gasification CHP (combined generation of electricity and heat) was considered commercial by this company, Figure 1.29 [54]. One of the most promising biomass gasification power plants network is the Renewable Energy Network Austria, which includes a gasification power plant that is supplying the town of Güssing with 2 MWe power and 4.5 MW of heat, from wood chips, since 2003, Figure 1.29 [55]. Figure 1.29: Harboøre (left) and Güssing power plants (right) [54, 55].
1. Introduction Study of a hybrid concentrating solar power plant for Portuguese conditions Page 38 The gasification technologies having been successfully demonstrated, however, they are still relatively expensive, and face economic and non-technical barriers when trying to be introduced into the national electricity generation networks. Most of the new projects address technical aspects of the gasification processes, but also the integration of gasification in existing projects, showing that the overall system offers better economic prospects. A different possibility is to use refuse-derived fuel (RDF) from municipal solid waste pellets (mainly plastics and biodegradable waste). The gasification of RDF is an interesting solution to solve the environmental impact of municipal solid waste. The Chianti RDF power plant is an example of a power plant using RDF gasification [56, 57]. The Greve plant is equipped with two 15 MWth TPS CFB gasifiers, each with a capacity of 100 t/d of RDF pellets. However, several operational problems occurred at Chianti, namely with gas cleaning and maintaining gas properties. The Chianti operational scheme is presented in Figure 1.30 [58]. Figure 1.30: Chianti power plant (left) and scheme (right) [58]. Locally there is also an interesting potential of biogas generated from a waste water treatment plant and landfills. The biogas generated from the landfill has different characteristics compared with the biogas generated by a wastewater anaerobic digester [59]. The Barlavento landfill (Algarve) receives 176 thousand tonnes (2011) of waste every year and generates biogas to run the 900 kWe power block (it is in progress the expansion to 1.6 MWe). The Sotavento landfill is the other landfill in Algarve region and started operation in 2011 with a 1.2 MWe power block, Figure 1.31 [60].
1. Introduction Study of a hybrid concentrating solar power plant for Portuguese conditions Page 39 Figure 1.31: Barlavento landfill (left) and Sotavento landfill (right) [60]. The biogas can also be generated from a wastewater treatment plant (WWTP). There are five main WWTP in the Algarve region: Vilamoura (140 000 inhabitants), Almargem (50 000 inhabitants), Faro (45 000 inhabitants), Olhão (35 000 inhabitants) and Boavista (25 000 inhabitants). All the Algarve WWTPs have no actual biogas valorisation. The larger national WWTP with biogas valorisation is the Frielas WWTP with a 2 MWe power block, which serves a population of 700 000 inhabitants, Figure 1.32 [61, 62]. Figure 1.32: Vilamoura (left) and Frielas (right) WWTPs [61, 62]. An alternative or supplement to biogas/syngas is the natural gas. Natural gas Rankine and combined cycle power blocks are an established technology, but the “renewable goal” would be lost. Commercial power plants running on natural gas normally use combined cycles up to several hundreds of MWe. All these base case power plants are analysed through the thesis. The integration of biomass from these sources and using several conversion technologies is also studied for the hybridization of the CRS. The innovative models present interesting results, which is going to be used to analyse the concept of CSP and biomass hybridization.
1. Introduction Study of a hybrid concentrating solar power plant for Portuguese conditions Page 40 1.5 Thesis structure, background and outputs The thesis has 7 chapters. A review of the state-of-the-art on CSP and biomass technologies, and their possible integration, is carried out in Chapter 1. A model for the CRS design was developed by the author, and is presented in Chapter 2. This new model was created in HFLCAL, Ebsilon and Excel, and optimized for the Portuguese conditions and solar only operation in Chapter 3. The work developed in the thesis started with the participation in a consortium that applied to a PIP call (SOLMASS project), in which FEUP was a scientific partner. Due to the promising local conditions, the SOLMASS project aimed to be the first hybrid Biomass/CRS power plant of its kind in the world. With this perspective, in Chapter 4, several biomass options were analysed for integration into the CRS steam cycle. The integration of biomass into the air cycle is presented in Chapter 5. Chemical generation, based in renewable resources and CSP, is also analysed, Chapter 6. Chapter 7 presents the main conclusions of the work and perspectives of future work.
1. Introduction Study of a hybrid concentrating solar power plant for Portuguese conditions Page 41 References [1] Schillings, C., Trieb, F., MED-CSP Study, Concentrating Solar Power for the Mediterranean Region, DLR, 2005. [2] Caracterização da Rede Nacional de Transporte para Efeitos de Acesso à Rede em 31 de Dezembro de 2008, REN, Março 2009. [3] Richter C., Teske S., Short R., Concentrating Solar Power Global Outlook 09, Greenpeace, Estela and SolarPaces, 2010. [4] Gordon, J., Solar energy: the state of the art : ISES position papers, International Solar Energy Society, 2001. [5] Johansson, B., Burnham L., Renewable energy: sources for fuels and electricity, Island Press, 1993. [6] Lovegrove, K., Luzzi A., Solar Thermal Power Systems, Encyclopedia of Physical Science and Technology, Volume 15, Academic Press, 2002. [7] PS 10 Final Technical Progress Report, Solúcar, 2006. [8] www.nrel.gov/csp/solarpaces/power_tower.cfm, accessed on 06-12-2010. [9] www.brightsourceenergy.com/, accessed on 06-12-2010. [10] www.esolar.com/, accessed on 06-12-2010. [11] http://www.weizmann.ac.il/weizsites/solarenergy/, accessed on 16-12-2012. [12] http://www.masdar.ac.ae, accessed on 16-12-2012. [13] http://technology4life.wordpress.com/2009/01/25/financing-secured-for-gemasolar- power-tower-from-torresol-energy/, accessed on 06-12-2010. [14] www.pratt-whitney.com/vgn-ext- templating/v/index.jsp?vgnextoid=f4d1202744f1a210VgnVCM1000004f62529fRCRD, accessed on 06-12-2010. [15] http://www.ka-muenchen.de/253+M52087573ab0.0.html, accessed on 06-12-2010. [16] SOLGATE - solar hybrid gas turbine electric power system Final Report, European Commission, 2005. [17] Giuliano, S., et al., Analysis of solar-thermal power plants with thermal energy storage and solar-hybrid operation strategy, Journal of Solar Energy Engineering, 133 (2011), p. 310071-310077.
1. Introduction Study of a hybrid concentrating solar power plant for Portuguese conditions Page 42 [18] Romero, M., Zarza, E., Handbook of Energy Efficiency and Renewal Energy, Chapter 21: Concentrating Solar Thermal Power, CRC Press, 2007. [19] Romero, M., Buck, R., Pacheco, James., An Update on Solar Central Receiver Systems, Projects, and Technologies, Journal of Solar Energy Engineering 124 (2002), p. 98 – 108. [20] Pitz-Paal, R., Dersch, J., Milow, B., ECOSTAR roadmap document, DLR, 2005. [21] Pacheco, J., Overview of recent results of the solar two test and evaluations program, Sandia Report, 1999. [22] www.torresolenergy.com/TORRESOL/gemasolar-plant/en, accessed on 06-12-2010. [23] Hoffschmidt, B., Téllez, F., Valverde, A., Fernández, J., Fernández, V., Performance Evaluation of the 200-kWth HiTRec-II Open Volumetric Air Receiver, Journal of Solar Energy Engineering 125 (2003), p. 87 – 94. [24] http://www.saint-gobain-solar-power.com/heat-receivers-hpr-13, accessed on 06-12- 2010. [25] Kearney, D., Kelly, B., Price, H., Thermal Storage Commercial Plant Design Study for a 2- Tank Indirect Molten Salt System, NREL report, 2006. [26] Tamme R., Laing D., Steinmann W., Advanced Thermal Energy Storage Technology for Parabolic Trough, Proceedings of the 2003 International Solar Energy Conference, 2003. [27] KAM technology catalogue. [28] www.saint-gobain-solar-power.com/solar-thermal-storage-norpro-9, accessed on 06- 12-2010. [29] Luz International, Phase-Change Thermal Energy Storage, NREL report, 1989. [30] Stine, W., Solar energy fundamentals and design, Wiley, 1985. [31] Siemens catalogue - www.energy.siemens.com/fi/pool/hq/power-generation/steamturbines/downloads/E50001-W410-A105-V1-4A00_Solarbroschuere.pdf, accessed on 06-12- 2010. [32] www.mandieselturbo.com/1013241/Press/Press-Releases/Press-Releases/First-steam- turbine-for-solar-power-station.html, accessed on 06-12-2010. [33] www.gepower.com, accessed on 06-12-2010. [34] http://www2.cnrs.fr/journal/3160.htm, accessed on 06-12-2013. [35] U.S. Department of Energy, Concentrating Solar Power Commercial Application Study: Reducing Water Consumption of Concentrating Solar Power Electricity Generation, Report to Congress, 2009. [36] www.nrel.gov/csp/troughnet/power_plant_systems.html, accessed on 06-12-2010.
1. Introduction Study of a hybrid concentrating solar power plant for Portuguese conditions Page 43 [37] www.nrel.gov/csp/solarpaces/, accessed on 06-12-2010. [38] Despacho n.º 18838/2009, Diário da República n.º 157, 2ª Série de 14 de Agosto de 2009 - http://dre.pt/pdf2sdip/2009/08/157000000/3313233132.pdf, accessed on 06-12-2010. [39] http://www.dgge.pt/, accessed on 06-12-2010. [40] CNADS, “Reflexão do CNADS sobre Energia e Sustentabilidade”, 2007. [41] www.renae.com.pt/_fich/22/060630DGGE.pdf, accessed on 06-12-2010. [42] Estratégia Nacional para a Energia, Resolução do Conselho de Ministros n.º 169/2005. [43] www.portugal.gov.pt/, accessed on 06-12-2010. [44] Estratégia Nacional para a Energia 2020, Resolução do Conselho de Ministros n.º 29/2010. [45] Coelho, B., Domingues, P., Oliveira, A.C., Mendes, A., "SOLMASS project - solarbiomass dual hybrid 4 MW CRS pilot plant", Solarpaces 2010, Perpignan, France, September 2010, paper 295. [46] http://www.logistica-florestal.pt, accessed on 16-12-2012. [47] Yakima County Public Works Solid Waste Division, Review of biomass fuels and technologies, R.W.Beck, 2003. [48] Neto, V., Potencial da Cogeração e Planeamento da Expansão do Sector Eléctrico, 2006. [49] Queirós, A., Produção de Metanol a partir de Biomassa Vegetal: Um novo processo integrado, Tese de Mestrado, FEUP, 2009. [50] http://www.oilgae.com/algae/algae/oil/extract/extract.html, accessed on 16-12-2012. [51] www.energiasrenovaveis.com, accessed on 06-12-2010. [52] Kayhanian, M., Tchobanoglous, G., Brown, R., Handbook of Energy Efficiency and Renewal Energy, Chapter 25: Biomass Conversion Processes For Energy Recovery, CRC Press, 2007. [53] Brandberg Å., Hjortsberg H., Sävbark B., Ekbom, T., Hjerpe, C., Landälv I. "BioMeeT, Planning of Biomass based Methanol Energy Combine at Trollhättan region". Final report by Trollhättan Municipality, Ecotraffic R&D AB and Nykomb Synergetics AB, Stockholm, 2000. [54] www.volund.dk/solutions_references/gasification_solutions, accessed on 06-12-2010. [55] www.renet.at/english/sites/guessing.php, accessed on 06-12-2010.
1. Introduction Study of a hybrid concentrating solar power plant for Portuguese conditions Page 44 [56] Granatstein, D.L., Case study on waste-fuelled gasification project greve in Chianti, Italy, IEA bioenergy agreement—task 36 report, 2003. [57] G. Barducci, G., Ulivieri, P., Pike, D.C., McDonald, N., Repetto, F., Cristo, F., The Greve in Chianti project, Renewable Energy 16 (1999) p. 1041–1044 [58] Morris, M., Waldheim, L., Energy recovery from solid waste fuels using advanced gasification technology, Waste Management 18 (1998) p. 557–564. [59] Persson, M., Jönsson, O., Wellinger, A., Biogas Upgrading to Vehicle Fuel Standards and Grid Injection, IEA Bioenergy report, 2006. [60] http://www.algar.com.pt, accessed on 10-07-2013. [61] http://www.aguasdoalgarve.pt, accessed on 10-07-2013. [62] http://www.simtejo.pt/, accessed on 10-07-2013.
Study of a hybrid concentrating solar power plant for Portuguese conditions Page 45 Chapter 2 Model of an atmospheric volumetric central receiver system (CRS)
2. Model of an atmospheric volumetric CRS Study of a hybrid concentrating solar power plant for Portuguese conditions Page 52 Figure 2.2: Image reflected by a heliostat with normal error distributions (left) and real image reflected by the same heliostat (right) [3]. 2.1.1.3 Errors for different incident angles: astigmatism The solar beam error, defined in equation 2.3, is valid only for design conditions - with incidence angle Ψ = 0 °. For different conditions the astigmatism effect of an image reflected by a concentrating surface has to be considered. = + # (2.4) A spherical or paraboloidal solar concentrator concentrates the parallel solar rays in a single focal point only for a specific incidence angle. For other incidence angles the solar radiation is concentrated between two lines (green and red, Figure 2.3). Two planes can be defined: the tangential plane and the sagittal plane. For f/d >> 1, the solar rays in the tangential plan (green) concentrate in the focal point at a f·cos Ψ distance. The rays in the perpendicular sagittal plane (red) are concentrated in a focal length of f/cosΨ, Figure 2.3 [3].
2. Model of an atmospheric volumetric CRS Study of a hybrid concentrating solar power plant for Portuguese conditions Page 53 Figure 2.3: Image reflected by a spherical concentrator [3]. For a given slant range - SLR distance, from a mirror to the surface - the image reflected has the following dimensions in the tangential and sagittal planes: $ =%& '() * −,-.Ψ&;1 =%& '() * ·,-.Ψ−1& (2.5) In HFLCAL, the image dimensions due to off-axis reflection can be described as superposition of astigmatism in the heliostat and in the facet: $ =% 45 −% 46 ·|1−,-.Ψ|+% 46 ·&,-.Ψ− '() * & (2.6) 1 =% 45 −% 46 ·|1−,-.Ψ|+% 46 & '() * ·,-.Ψ−1& (2.7) where % 45 and % 46 are the diameters of the heliostat and facets and 8 is the focal length of the facets. In HFLCAL, the root-mean-square (RMS) of the reflected solar image (in the tangential and sagittal plane) is treated as uniform distribution and is incorporated as additional error to the reflected solar beam with half width of # .
2. Model of an atmospheric volumetric CRS Study of a hybrid concentrating solar power plant for Portuguese conditions Page 54 # = 9 :;< =>? @ A∙'() (2.8) The astigmatism error is mainly associated with the heliostat size (for large incident angles) while for low incident angles the facet size is dominant. For heliostats with astigmatism correction HFLCAL adds the possibility of introducing a correction factor. 2.1.2 Annual simulation The performance of a heliostat is dependent on several factors, which can be divided into three major groups: • Regardless of the time and position of the heliostat: reflectivity; • Dependent only on position: atmospheric attenuation; • Depending on the time and position: cosine factor, blocking and shading, interception. Considering these factors, the energy incident into a certain surface (e.g. solar receiver) from a heliostat with an area for a specific instant (t), under a direct normal irradianceCDEF, is given by the following equation [3]: GH,J,F=CDEF∙ ∙K * ∙K H,J∙K ! H,J,F∙K & H,J,F∙K 4 H,J,F (2.9) From the sum of all heliostats the energy supplied by the field of heliostats to the surface of the solar receiver can obtained for a given time period by: G *M F=∑G H,J,F (2.10) HFLCAL uses about 100 periods of time, typically the daylight hours of the 21 st day of each month, to calculate the annual performance, which depends on the location selected for the solar plant (e.g. latitude and height above sea level). 2.1.3 Main factors that influence the solar field layout The solar field annual efficiency depends on a large number of factors: cosine factor, shading, blocking, atmospheric attenuation and spillage. The solar shape, mirror curvature and defects or imperfections in the mirror surface are also responsible for the final size of the focal point for a given point in time. All these effects overlapped form the solar image incident on the receiver.
2. Model of an atmospheric volumetric CRS Study of a hybrid concentrating solar power plant for Portuguese conditions Page 55 2.1.3.1 Reflectivity The reflective material is an important for the design and efficiency of the solar field. Currently various materials are used and several companies produce specific and optimized reflective materials for application in solar-thermal power plants (cylinder-parabolic, Fresnel, Stirling disks or CRS), e.g. Saint-Gobain, Flabeg and Guardian. Normally, reflective materials are divided according to the position of the reflective surface: 1 st surface, 2 nd surface or multi surface. In the case of CRS, the material typically used is a silver film on thick glass support. This is a 2 nd surface mirror and has a longer operational experience under real conditions; typically keeps its characteristics for periods of 25 years with only slight breaks in reflectivity. There are also innovative 2 nd surface materials such as the silver film on thin glass, which has good features and price, yet still being commercially validated (mainly to maintain the characteristics for periods similar to the silver film in thick glass). 1 st surface materials such as the aluminium film polished on a metallic support may enable combining the structure of the heliostat with its reflecting surface, bringing a possible cost reduction; currently these materials have low reflectivity and durability, requiring frequent repositions of the reflecting surface to maintain the overall efficiency of the solar field. An overview of the features and costs of reflective materials is presented in Table 2.2. Table 2.2: Characteristics of the mirrors available on the market. Material Type Reflectivity Thickness Durability Price Aluminium film on metallic support 1 st surface 0.83 – 0.87 0.3 to 1 mm 5 years 15 €/m 2 Silver film in thick glass 2 nd surface 0.935 3.5 to 5 mm over 20 years 35 - 65 €/ m 2 Silver film in thin glass 2 nd surface 0.93 0.4 to 1.2 mm 20 years 15 €/m 2 Prices are dependent on quantities and also vary substantially with the time of acquisition [4].
2. Model of an atmospheric volumetric CRS Study of a hybrid concentrating solar power plant for Portuguese conditions Page 56 2.1.3.2 Cosine factor The cosine factor takes into account the reduction of the surface that effectively reflects the solar radiauon at incident angles Ψ ≠ 0. It is one of the most important factors for an efficient solar field. The cosine factor varies with solar position and the position of each heliostat. It is assumed that the heliostats have a correct tracking, and as such, their cosine factor is defined by the cosine of the angle formed by the incident solar beam and the surface normal vector of the heliostat (Ψ): K ! =cosΨ (2.11) The heliostat tracking mechanism follows the sun in a way that its normal vector bisects the angle between the solar ray, and a vector pointing to the focal point on the receiver. This can result in a clear decrease in the effective heliostat reflective area, Figure 2.4. Figure 2.4: Cosine factor (left) [6] and effect on the solar field (right). For this reason, a heliostat in the northern hemisphere, when placed in the south area (heliostat B – Figure 2.4) has a lower efficiency than the heliostats placed in the north area, which face the sun (heliostat A – Figure 2.4). In a similar analysis, in the morning the heliostats placed west of the tower have higher efficiency than the heliostats placed east, and during the afternoon the opposite occurs. 2.1.3.3 Blocking and shading For the calculation of heliostat blocking and shading losses, HFLCAL projects the image of the central heliostat in the surrounding group of heliostats. This analysis is carried out in two different directions: between the sun and the heliostat to calculate the shading, and between the heliostat and its focal point (solar receiver) to calculate the blocking. Shadowing occurs when a heliostat, or part of it, is shadowed by one or more neighbours. Blocking occurs when a part of the heliostat sees its solar reflection in the receiver blocked by their neighbours
2. Model of an atmospheric volumetric CRS Study of a hybrid concentrating solar power plant for Portuguese conditions Page 57 [5, 6]. Figure 2.5 represents the losses from blocking and shading and the heliostats in the solar field usually most affected by these losses. Figure 2.5: Blocking and shading losses (left) [6] and effect on the solar field (right). The blocking and shading losses are a function of the spacing between heliostats, tower height and solar angles, which are dependent on the respective time of day and position of the heliostats. HFLCAL has different distribution patterns to minimize these losses. The total non-shaded area is calculated and the ratio between this area and the total area is found to obtain the blocking and shading efficiency. Other simulation tools, such as Helios and Delsol2, also use this approach, positioning the heliostats more compactly near the tower and with more spacing far from the tower, so the blocking and shading losses are minimized. An indicator of the solar field layout is its heliostat field density. This is the ratio between the mirror area and the total area of the field or certain zone. It can be calculated by Equation 2.12, where C is the mirror density (ratio of heliostat mirror area and heliostat total area): R *M = ∙S TUV? ∙W XYZ[\ ·W ]^_`[\ ∙abcdefghiejkglmnmlkamieokejp *M (2.12) 2.1.3.4 Atmospheric attenuation The atmospheric attenuation of the rays reflected by the heliostats is dependent on the distance between the receiver and the heliostat, the slant range - SLR. K qr! =0.99321−0.0001176∙yz{+1.97×10 }~ ∙yz{ (2.13) For slant ranges of more than 1 km (SLR > 1000 m): K qr! = }•.••••€∙'() (2.14)
2. Model of an atmospheric volumetric CRS Study of a hybrid concentrating solar power plant for Portuguese conditions Page 58 The models used by HFLCAL are in agreement with experimental data collected by NREL during operation of the solar plant Solar Two in Daggett, California, USA [7] (Figure 2.6), for clear sky models (HFLCAL and Solar Two - 23 km). For the most common slant range of commercial solar fields (up to 1.5 km), HFLCAL variations from measured data are 2 to 3%. Although they are specific to a location in California, the NREL values may be used for other similar locations. However, significant changes are recorded for locations with higher values of aerosols at ground level (commonly referred as visibility) and for locations with different altitudes. Figure 2.6: Impact of atmospheric attenuation on the efficiency of a solar field for HFLCAL and measured data from Solar Two (lines are for readability). For hazy days (Solar Two 5 km) the decrease in efficiency (with increasing distance to the solar tower) is more intense - losses closer to 50 % at 2.5 km. Commercial heliostat fields can reach distances of about 1.5 km and atmospheric attenuation losses become significant (≈ 15 %), even considering models of clear sky, Figure 2.7. Atmospheric attenuation losses, associated with interception losses, mean that there is a technical/economic limit for the field area (and ultimately for the CRS installed power). 2.1.3.5 Intercepted radiation The flow distribution of the solar beams reflected from each heliostat must be integrated over the receiver area to obtain the effective incident energy in the receiver at a certain point in time. In HFLCAL each solar image reflected by a heliostat is described by Equation 2.1 and the reflected image intercepted by the receiver is expressed by the following equation:
2. Model of an atmospheric volumetric CRS Study of a hybrid concentrating solar power plant for Portuguese conditions Page 59 K !q = <V<•‚ ∬ } „=… <V<•‚ !† %H%J (2.15) The beam dispersion width is calculated according to Equation 2.4. In the case of a circular aperture with the heliostat aiming at the centre this integration is solved analytically. In all other cases numerical integration routines are used. The spillage factor (or the ratio of the reflected image that cannot intercept the receiver surface) is a function of the quality of heliostat solar tracking, uniformity of the reflective surface, the shape of the sun, environmental factors such as wind speed, heliostat foundation, controller and tracking algorithm used, Figure 2.7 [5, 6]. Figure 2.7: Effect of atmospheric attenuation in the solar field (left) spillage (centre) [6] and the respective effect of the solar field (right). 2.1.4 Distribution/optimization of the solar field 2.1.4.1 Heliostats The starting point for solar field optimization is the definition of the individual heliostat dimensions and the type of heliostat. The most common heliostat configuration is rectangular, with square or rectangular mirrored facets, Figure 2.8. Each facet and heliostat reflects a circular image [3].
2. Model of an atmospheric volumetric CRS Study of a hybrid concentrating solar power plant for Portuguese conditions Page 60 Figure 2.8: Heliostat scheme and dimension [3]. ALV and ALH are, respectively, the vertical and horizontal dimension of the heliostat and Z is the height above the ground up to the centre of the heliostat; HFCAL also requires to define the FMIR and FFAC, respectively the heliostat mirror surface and its facet areas; EFELD and SIG are respectively the annual average reflectivity and total error of the solar beam reflected by the mirror surface. It is also possible to select two options: ideal focus distance - which considers all facets focal points match the slant range; and heliostats with correction of their astigmatism. The project Helios3S designed a heliostat for application in SOLMASS CRS. The process of optimization of the heliostats size depends on the configuration and characteristics of the receiver and maximizes the energy generated by the plant, but also optimizing the solar field cost. Regarding various commercial approaches to the configuration and dimensions of a heliostat field (Chapter 1), the selected approach was to create a medium-sized heliostat (Helios3S) with the characteristics presented in Table 2.3.
2. Model of an atmospheric volumetric CRS Study of a hybrid concentrating solar power plant for Portuguese conditions Page 61 Table 2.3: Characteristics of the Helios3S heliostat. Characteristic Value Characteristic Value ALV 7.76 m FFAC 5.00 m 2 ALH 7.76 m EFELD 87 % Z 4.88 m SIG 3.60 mrad FMIR 60.32 m 2 2.1.4.2 Solar field layout HFLCAL uses an algorithm to distribute the heliostats throughout the solar field, selecting afterwards the most efficient heliostats to generate the thermal energy required by the receiver. The distribution of heliostats can be based on three algorithms: bilinear expanded, bilinear with spacing and slip planes. Each model has different characteristics and may be more advantageous in terms of efficiencies, depending on the type of receiver and the design power, Figure 2.9. Figure 2.9: Typical heliostat layout for different algorithms: bilinear expanded (left), bilinear with spacing (centre) and slip planes (right).
2. Model of an atmospheric volumetric CRS Study of a hybrid concentrating solar power plant for Portuguese conditions Page 68 The receiver design conditions and its operating temperatures and efficiencies were set in combination with the power cycle model, which was designed in Ebsilon. The receiver performance was interpolated from Figure 2.15. 2.1.4.5 Receiver solar flux The mapping of solar fluxes incident on the receiver is essential to analyze the performance of the solar receiver. A correct distribution of the solar flux over the entire area of the receiver allows an increase of the receiver efficiency and longevity. Too high fluxes may cause structural deformations in the receiver cups and structure. Typically, for SiC receivers, currently used in commercial CRS, peak flows should be max. 1100 kW/m 2 . For large power this is only possible by defining a strategy with several focal points, each focusing on solar images of a certain group of heliostats, usually using groups of heliostats closer to the solar tower to focus on points nearer to the receiver boundaries and farther heliostats (with larger solar image) to cover the points of focus in the centre of the receiver. HFLCAL allows using different strategies and focal points in the receiver, Figure 2.16. Figure 2.16: Strategies for solar field focusing: central point (left) and several points on a centre line (right). The central point focussing strategy is used in various micro power plants or for thermochemical solar applications, normally using circular receivers or reactors (e.g. quartz window). For a 4 MWe power plant this strategy would reach very high concentrations in the centre of the receiver (Figure 2.16 - left), and so it is advisable to use more focal points in the receiver. One possible strategy is to use three focal points (similar to what is used in Jülich power plant); however, the level of concentration reaches more than 1300 kW/m 2 (Figure 2.16 - right) and should be avoided. HFLCAL has a distribution algorithm for the heliostat focal
2. Model of an atmospheric volumetric CRS Study of a hybrid concentrating solar power plant for Portuguese conditions Page 69 points into receiver, according to their image. It is also possible to optimize the focal point distribution, which is itself an iterative process. The goal is to limit the maximum incident solar peak flux and get a uniform flux profile throughout the area of the receiver. A good solar flux distribution may allow using a smaller receiver and consequently lower heat losses and cost. Nevertheless, a decrease in the solar receiver area, can lead to increased spillage losses and a consequent reduction in receiver efficiency and output power. 2.1.4.6 Solar field optimization The process of solar field optimization has the utmost importance for the overall lifetime performance of the solar power plant. HFLCAL has algorithms for optimization based on two principles: maximizing the performance of the power plant or minimizing the cost associated with it. It is necessary to estimate several initial values and a first performance of the power plant is calculated; subsequently, each iteration step is defined and new variable values (one by one) are defined and the new solar field configuration performance is calculated; the process is repeated, step by step, until a relative or absolute maximum is obtained. HFLCAL contains three optimization algorithms: iso-scan; genetic and Powell method. Each algorithm has specific characteristics and can be used at different stages in the optimization process. The iso-scan algorithm (all possible combinations) permits a scan of all selected parameters, combining them with each other, for a selected range of values (usually wide) in order to obtain an idea of what could be the values of interest to start the detailed optimization process. The computation time of this algorithm increases rapidly with the number of parameters selected and the range of values chosen. The genetic algorithm initiates a random collection of configurations with different values for the selected parameters and in accordance with the algorithm starts an internal convergence process for obtaining the maximum overall optimization function for the parameters selected. The algorithm is efficient in getting the maximum, by phasing elimination of not maximum function values; however, this algorithm needs further optimization processes to obtain the specific values of the parameters that the algorithm estimates. This optimization can be done using the Powell method. This method gives only a good approximation (with reasonable times) when the initial values of the parameters are chosen close to the maximum values. Since this is a directional method, it will progressively increase values of the selected parameters until a maximum is found, and the selected parameters should be reviewed as they may exceed the permissible physical values.
2. Model of an atmospheric volumetric CRS Study of a hybrid concentrating solar power plant for Portuguese conditions Page 70 Several optimization algorithms were used to find the best solar field layout. The process is limited by equipment operability and physical barriers (when a minimization of costs is selected) and by the cost of the equipment (when a maximization of the performance is selected). The whole process is repeated in order to maximize the performance and minimize cost of the solar field according to all design parameters. The optimization process is briefly described in the decision tree of Figure 2.17. Figure 2.17: Solar field optimization strategy.
2. Model of an atmospheric volumetric CRS Study of a hybrid concentrating solar power plant for Portuguese conditions Page 71 2.2 Power circuit model 2.2.1 Ebsilon Professional Ebsilon Professional software simulates thermodynamic cycle processes and is used for engineering, designing, and optimizing power plants [10]. During the design process, it can be used to identify optimal cycles and evaluate various options and configurations. It can also be used during plant operation to evaluate losses and suggest improvements. Ebsilon uses a graphical interface that allows the models to be built using single components, groups of components, sub systems or complete systems. Ebsilon has an extensive component library (e.g. heat-exchangers, boilers, pumps, steam and gas turbines, fans, etc) and different fluid media property libraries and tables [10]. It has a design mode and an off-design mode; the power plant design mode sizes the equipments for nominal operational conditions. Off design operation, necessary for annual simulation, is usually defined by characteristic lines. Conventional (natural gas or coal) power plants usually operate in close to nominal conditions almost all year round. This is not the case of a CRS, which is more dynamic (because solar irradiance is not constant), and because of that, several operational profiles are necessary, Figure 2.18. Figure 2.18: Ebsilon structure.
2. Model of an atmospheric volumetric CRS Study of a hybrid concentrating solar power plant for Portuguese conditions Page 72 2.2.2 Power circuit optimization approach Solar power has several peculiarities that influence the power circuit equipment selection and how the power plants are designed to incorporate it. As mentioned in Chapter 1, for commercial systems three types of CSP thermal cycles may be considered: Rankine, Stirling and Brayton cycles. The choice of thermal power cycle depends on the type of technology selected. For CRS (as in the case of LF and PT) a Rankine or Brayton cycle is usually selected, which may unfold in several power circuit configurations and operational conditions. To determine the best power circuit configuration for the CRS, several variables are of great importance: e.g. the electrical power to inject into the network (4 MWe), the available solar irradiance (Faro, typical meteorological year, DNI of 2183 kWh·m -2 ·yr -1 ), power block operational conditions, solar multiple, storage capacity and control strategy. The final variable definition is an optimization process including the entire solar plant: solar field layout, solar receiver configuration, thermal cycles used among others. This optimization process should be carried out in accordance with technical and economic criteria so that the CRS has the highest possible performance with the lowest possible costs. In parallel to this process it is vital to validate the solar power plant components and models with commercial equipment. This is an extremely laborious process since there are often conflicting interests among different equipment manufacturers, as well as difficulty in disseminating performance characteristics for reasons of confidentiality. Moreover, the number of suppliers is reduced and typically composed of an oligopoly of large multinationals (SIEMENS, MAN TURBO; ALSTOM, etc.), who impose signed confidentiality agreements to transfer only partial information. There are components, such as the solar receiver and the heliostat field, that are still in early commercial phase and as such the information on their performance is reduced and extremely confidential. In the following chapter is presented the design process and optimization of the power circuit of a CRS of 4 MWe, to be implemented in the region of Algarve (Faro). As shown in Figure 2.19, the design of CRS is typically done from downstream to upstream. The first variable to be set is typically the peak power to be injected into the grid. In the case of the considered CRS power plant, the installed power is 4 MWe and a power block for this power should be found. The detailed definition of the power circuit equipments and their operational conditions is not only dependent on the installed power but also the thermal
2. Model of an atmospheric volumetric CRS Study of a hybrid concentrating solar power plant for Portuguese conditions Page 73 cycle to be applied, which is influenced by the receiver technology, the solar field, local conditions and control strategy. Figure 2.19: Power circuit optimization strategy.
2. Model of an atmospheric volumetric CRS Study of a hybrid concentrating solar power plant for Portuguese conditions Page 74 Thermal engines and machines have a theoretical maximum efficiency when converting energy to work: the Carnot efficiency. During this conversion, entropy accumulates in the system, and is removed by dissipating heat. The power cycle receives thermal energy at high temperature, converts some of that energy into mechanical work, and rejects the remaining at a lower temperature. The thermal efficiency of an engine operating under Carnot cycle is defined as: K ¯ = §"#M rq =1− ¥ V¦< ¥ Uœ (2.24) where T out and T in are the absolute temperatures at which heat is added and rejected by the Carnot heat engine, respectively. The thermal efficiency of an engine is therefore proportional to the ratio between the input temperature of the cycle and the temperature of heat rejection. The wider the difference in temperatures, the more efficient the conversion of heat into work by the Carnot engine is. Because it is a limiting scenario, real engines operate in other cycles than the Carnot cycle. Those cycles operate at lower efficiencies than the Carnot cycle. However, the effect of temperature on the efficiency is still valid for real engines, Equation 2.25. The variation in efficiency of a real engine can then be represented by: K ) =° ) ×K ¯ (2.25) where K Real represents the fraction of the Carnot efficiency reached by the real engine and K ¯ the equivalent Carnot cycle efficiency. Equation 2.24 indicates that power circuit efficiency increases with the increase in input temperature T in . The use of very high temperatures is limited by the materials applied and the costs associated to the power block. Contrarily, the solar receiver has higher efficiencies for lower operational temperatures, mainly due to the radiation losses, Equation 2.26 and 2.27 [8]. K )!† = ± V¦< ² ¬,• ¡ • (2.26) K )!† =K ³q − ´ ‚ ¥ }¥ • ² ¬,• ¯) µ − ª ¶ :¥ « }¥ •« @ ² ¬,• ¯) µ (2.27)
2. Model of an atmospheric volumetric CRS Study of a hybrid concentrating solar power plant for Portuguese conditions Page 75 where ˆ is the receiver area (m 2 ); ·{ # is the geometrical concentration ratio; E , is the radiation flux into the receiver (W·m -2 ); ¸ is the receiver output heat (W); • is the atmospheric temperature (K); • is the receiver operational temperature (K); ‹ is the receiver heat transfer coefficient (W·m -2 ·K -1 ); ¹ is the receiver emissivity and º is the Stefan-Boltzmann constant (5.6696 × 10 -8 W·m -2 ·K -4 ). According to Stine et at. [8] the overall efficiency of the system is well approximated by the product of the real machine efficiency and effectiveness of the collector. The analysis of the overall efficiency versus the operating temperature curve allows finding an operating temperature • !,“ that maximizes system efficiency, Equation 2.28 and 2.29 [8] which, applied to the selected receiver, results in Figure 2.20. 4· ¼ ½ “¾ −3· ¼ ½ “A +· ½ “ =· +· +· ¼ (2.28) ½ “ = ¥ ›¢,T•„ ¥ • ;· =K ³q ;· = ´ ‚ ¥ • ² ¬,• ¯) µ ;· ¼ = ª ¶ ¥ •« ² ¬,• ¯) µ (2.29) Figure 2.20: Definition of the best power plant operational temperature range. For CRS the range of ideal operating temperature is between 600 and 900 °C, Figure 2.20. For this temperature range, the CRS efficiency is maximized, although it may be extended from 500 to 1100 °C with only a small drop in efficiency. 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% 100 300 500 700 900 1100 1300 1500 Efficiêncy % Temperature (°C) η Power block η Receiver η CRS
Study of a hybrid concentrating solar power plant for Portuguese conditions 2.2.3 Steam cycle 2.2.4.1 Steam turbine Turbines are the most commonly used equipment for expansion in solar Rankine cycles. Typically, the efficiency of these devices is measured in relation to the ideal adiabatic and reversible expansion. An ideal turbine operates at constant entropy, unlike c turbines, which have higher entropy at steam outlet than at inlet. As already mentioned, there are several manufacturers of steam turbines, and each is in the process of developing new equipments to use in CSP plants. The process of modelling a c detail can be extremely complex. Ebsilon has a simplified steam turbine model that was used, adding some characteristic lines for off Figure 2.21 : In Ebsilon the turbine number of stages and extraction can be defined. In the case of the last stage, the pressure must be set by the user to indicate the existence of back pressure, or not, in the turbine system. Regarding turbine losses (at the entrance and exit), they can be calculated using the parameter QLOSSM regardless of the load. Alternatively, two characteristic lines can be defined (CKIN1 and CKIN2), representing losses in the entrance and exit of the turbine stage, respectively, when the operational mode (FSPEC) is set to "Total isentropic ef A turbine manufacturer (SIEMENS) was contacted to determine which would be the most suitable steam turbine for application in the specific CRS, and that would allow a future adaptation of a biomass module running in parallel/series to obtain the 2. Model of an atmospheric volumetric CRS Study of a hybrid concentrating solar power plant for Portuguese conditions Turbines are the most commonly used equipment for expansion in solar Rankine cycles. Typically, the efficiency of these devices is measured in relation to the ideal adiabatic and reversible expansion. An ideal turbine operates at constant entropy, unlike c turbines, which have higher entropy at steam outlet than at inlet. As already mentioned, there are several manufacturers of steam turbines, and each is in the process of developing new equipments to use in CSP plants. The process of modelling a c ommercial steam turbine in detail can be extremely complex. Ebsilon has a simplified steam turbine model that was used, adding some characteristic lines for off -design operation, Figure 2.21. : Ebsilon steam turbine graphical representation . In Ebsilon the turbine number of stages and extraction can be defined. In the case of the pressure must be set by the user to indicate the existence of back pressure, or not, in the turbine system. Regarding turbine losses (at the entrance and exit), they can be calculated using the parameter QLOSSM - which is a constant parameter of mechan regardless of the load. Alternatively, two characteristic lines can be defined (CKIN1 and CKIN2), representing losses in the entrance and exit of the turbine stage, respectively, when the operational mode (FSPEC) is set to "Total isentropic ef ficiency". A turbine manufacturer (SIEMENS) was contacted to determine which would be the most suitable steam turbine for application in the specific CRS, and that would allow a future adaptation of a biomass module running in parallel/series to obtain the Model of an atmospheric volumetric CRS Page 76 Turbines are the most commonly used equipment for expansion in solar Rankine cycles. Typically, the efficiency of these devices is measured in relation to the ideal adiabatic and reversible expansion. An ideal turbine operates at constant entropy, unlike c ommercial turbines, which have higher entropy at steam outlet than at inlet. As already mentioned, there are several manufacturers of steam turbines, and each is in the process of developing new ommercial steam turbine in detail can be extremely complex. Ebsilon has a simplified steam turbine model that was used, . In Ebsilon the turbine number of stages and extraction can be defined. In the case of the pressure must be set by the user to indicate the existence of back pressure, or not, in the turbine system. Regarding turbine losses (at the entrance and exit), they can be which is a constant parameter of mechan ical losses, regardless of the load. Alternatively, two characteristic lines can be defined (CKIN1 and CKIN2), representing losses in the entrance and exit of the turbine stage, respectively, when the A turbine manufacturer (SIEMENS) was contacted to determine which would be the most suitable steam turbine for application in the specific CRS, and that would allow a future adaptation of a biomass module running in parallel/series to obtain the installed power
2. Model of an atmospheric volumetric CRS Study of a hybrid concentrating solar power plant for Portuguese conditions Page 77 granted by the PIP (SOLMASS project). The suggested turbine was the SST-110 Model (Figure 2.22), characterized for its quick starts without preheating and by a "twin" structure, which is based on two individual "twins" that allow a great versatility of operation; an advantage in terms of increased hybridization possibilities and part load operation. Figure 2.22: SIEMENS steam turbine SST-110 Model [11]. Alternatively, a concept of three stages composed of three TSS-060 modules has been proposed [12], which also increases the versatility of integration, but showed a 10% higher cost compared to the SST-110 solution. Both options can be tuned to operate under different conditions: so, a large spectrum of options was requested to the manufacturer. Unfortunately, only a few configurations were given by the manufacturer and theoretical models were developed to study different possibilities/concepts to be applied in the power plant. These models were subsequently adjusted to agree in terms of performance with commercial data from the manufacturer. In the case of power plant construction the models should be readjusted to exactly match the commercial solution. This was also the methodology in the case of the electric generator and the heat recovery steam generator. 2.2.4.2 Heat recovery steam generator (HRSG) HRSGs are usually classified according to firing, layout and configuration. One of the ways to classify the HRSG is based on the use of auxiliary power: with or without duct burner. Some HRSGs only use exhaust gases (e.g. from a gas turbine or other process as a power source) and have their performance affected by this upstream equipment: e.g. in the case of a gas turbine part load operation, the HRSG steam production can be reduced, affecting its operation. To avoid such situations, a duct burner or supplementary firing equipment can be added. Another way of classifying the HRSG is the gas flow path: vertical or horizontal. In terms of performance and cost, these are equivalent systems and its use is only dependent on the type of manufacturer or customer preference regarding the layout. A different way to classify
2. Model of an atmospheric volumetric CRS Study of a hybrid concentrating solar power plant for Portuguese conditions Page 84 2.2.4 Air cycle 2.2.5.1 Receiver and solar multiple A particularity of solar thermal applications is that usually the energy is given by an external source (solar receiver) that heats a heat transfer fluid (HTF), as opposed to internal combustion engines where the potential energy is already intrinsic to the fluid. Therefore, it is necessary to choose the HTF, which can be pumped directly to the receiver or by incorporating an additional intermediate HTF between the receiver and power circuit. There are advantages and disadvantages to both. The incorporation of an intermediate HTF results in the need for another pump, a heat exchanger and a second HTF (results in increased system complexity, but often will reduce the size of the receiver due and the cost of high pressure piping). Pumping the working fluid directly to the power cycle can make the system difficult to control, especially in transient periods, because for Rankine cycle systems, preheating, evaporation, and overheat would occur in the receiver. Despite that, the concept is quite simple, may work more efficiently and with less initial investment, since less components are necessary. Receiver flux design is another variable to consider. Tubular receivers are difficult to operate at incident fluxes above 600 kW/m 2 (peak) [21], because the tubes may not resist the continuous thermal and mechanical stresses. In volumetric receivers (applied in atmospheric air technology), highly porous structures operate as convective heat exchangers absorbing the concentrated solar radiation. Air is forced through the porous structure (cooling the structure) and is heated by convective heat transfer, allowing incident fluxes larger than 1000 kW/m 2 (peak) [21]. Table 2.6 presents the operating temperatures and flux ranges of the available CRS solar receiver technologies [21]. Table 2.6: Operating temperatures and flux ranges of CRS solar receivers [21]. Fluid Tubular receiver Volumetric Air Water/Steam Liquid Sodium Molten Salt (nitrates) Average Flux (kW/m 2 ) 100-300 400-500 400-500 500-600 Peak Flux (kW/m 2 ) 400-600 140-2500 700-800 800-1000 Fluid outlet temperature (°C) 490-525 540 540-565 700 – 800 (>800)
2. Model of an atmospheric volumetric CRS Study of a hybrid concentrating solar power plant for Portuguese conditions Page 85 The receiver design is dependent on the thermal power necessary for the power block and the excess of power necessary to fill the storage device. This ratio is defined as the solar multiple: the ratio between the thermal power generated by the collector system (solar field and receiver), at the design point (DP) and the thermal power required by the power block at nominal conditions, Equation 2.31. yà S¤ = ± È›¢,¨© ± ©¶,¨© (2.31) In solar-only power plants the solar multiple is always greater than one, so the full load power block operation is not confined only to clear sky solar conditions. The increase in solar multiple also represents a higher capital investment (CAPEX), larger solar field, land area and receiver costs. Also the increase in solar multiple represents an increase in the solar field intercept power, Figure 2.29. Figure 2.29: Typical summer day history of available power from the solar field and generated electricity as a function of different solar multiples (lines are for readability). The power plant with 1.75 solar multiple (SM1.75_7S, Figure 2.29) collects more solar energy than the power plant with solar multiple of 1.25 (SM1.25_7S, Figure 2.29) and, as result, the period of power block full load operation is extended. This is only possible using thermal energy storage (TES); otherwise the energy exceeding maximal power block input must be dumped. 0 2 4 6 8 10 12 14 16 18 20 0 10 20 30 40 50 60 70 80 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 Enet (MWe) QDNI (MWth) Hour QDNI SM1.25_7S QDNI SM1.75_7S Enet SM1.25_7S Enet SM1.75_7S
2. Model of an atmospheric volumetric CRS Study of a hybrid concentrating solar power plant for Portuguese conditions Page 86 2.2.5.2 Thermal energy storage – storage size With TES it is possible to decouple power generation from the solar resource. Electricity generation can occur without solar resource and/or on demand by the electricity network, Figure 2.29. The TES unit provides an affordable and efficient solution to dispatchable CSP electricity generation. It is thus possible to adapt the electricity generation profile to the national electricity demand, decreasing the need of backup and stand-by power. The Portuguese electricity consumption during a typical summer day is presented on Figure 2.30. Figure 2.30: Scenario for CSP potential and Portuguese electricity consumption - REN typical day [22] (lines are for readability). A possible way to supply the electricity demand with CSP is by using large thermal energy storage that allows operating almost on a 24-hour daily base, Figure 2.30. As this is a costly solution, an alternative is to use the CSP hybridization potential. But even using only CSP it is possible to respond to the Portuguese network consumption day peak (7:00 to 24:00) with 625 CRSs (4 MW e each with SM of 1.75 and 7 hour storage – total 2.5 GW installed power), generating 10.2 TWh of electricity per year. This is far from the 142 TWh economic potential referred by the MED-CSP study [23] for CSP power plants in Portugal. Although this scenario is not probable to occur in the short term, CSP technologies can provide base load power, either in solar only mode or in more cost efficient hybrid solutions. Even in these hybrid solutions the storage unit confers stability and reliability to the electricity generation, preventing component failure and improving power plant performance. 0 1000 2000 3000 4000 5000 6000 7000 8000 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 Power (MW) Hour REN Typical day base load + Enet SM1.75_7S (625xCRS) Base load Enet SM1.75_7S (625xCRS) Other
2. Model of an atmospheric volumetric CRS Study of a hybrid concentrating solar power plant for Portuguese conditions Page 87 As in the case of atmospheric air volumetric CRS the HTF is atmospheric air, and, because air has very low energy density and conductivity, direct active heat storage is not a good solution. The technology applied in the Jülich solar tower is a regenerator-type storage (passive storage), where a gaseous heat transfer fluid is in direct contact with a solid storage medium and exchanges heat as it flows along a path through the storage medium [24]. The storage device is a rectangular housing of 7 x 7 x 6 m 3 (total volume of 120 m 3 ), divided into four chambers of identical size, filled with a ceramic storage material, and connected in parallel. The storage system operates between 120 and 680 °C and has a capacity of almost 9 MWh. The total heat loss in a 24 h period is 930 kWh (fully charged storage) with a pressure drop of 15 mbar [24]. There are some limitations on the number of equivalent hours of storage (between 3 and 6 hours), due to technological and economic reasons [25]. The optimal storage capacity depends on the solar multiple and control strategy. The control strategy is the power plant operational strategy and varies during the day, e.g. according to network special needs, contract with the electricity purchase entity, feed-in tariff, technologies used in the power plant and available staff. Several countries have premium feedin electricity tariffs for producers during the day, while others have fixed feed-in tariffs or forecast obligations [26]. In the Portuguese case, the power plant operator can choose if he wants to receive the same remuneration regardless of the time of day, or a higher tariff for electricity generated during the day than during the night. In the last case, the amount of electricity generated between 8:00 and 22:00 during wintertime and 9:00 and 23:00 during summertime is multiplied by 1.25 and the rest of the electricity is multiplied by 0.6 [26]. As the feed-in tariff calculation formula is complex, the regulators release the value for each project or call; in the case of the latest CSP call, the feed-in tariff is 0.273 €/kWh [26]. 2.2.5.3 Ebsilon model for the air cycle The receiver model was based on the 3D interpolation of measured data from Jülich solar tower, provided by DLR on Figure 2.15. The receiver input is the solar intercept power given by HFLCAL which heats a cold air stream (60 % recycled and 40 % fresh air) up to a temperature of 680 °C (at design conditions) with a flow rate defined by the discharge and charge controllers. Ambient conditions are based on Meteonorm hourly data. The hot air stream is forwarded to the storage device and the HRSG, Figure 2.31.
Study of a hybrid concentrating solar power plant for Portuguese conditions Figure 2. 31 The quantity of energy that is forwarded to the storage is defined by the solar multiple and the control strategy. At design conditions the HRSG is supplying the power block required energy to generate the 4 MWe net and the excess is stored. The ceramic st for large contact surfaces so heat transfer is optimized. The air can flow through the storage in both directions (charging and discharging) but the storage temperature varies and discharging. Due to Ebsilon restrictions t hot tank Figure 2.31. However in the power plant there is a single storage device with the temperature profiles presented in 2. Model of an atmospheric volumetric CRS Study of a hybrid concentrating solar power plant for Portuguese conditions 31 : Ebsilon model for the receiver and storage. The quantity of energy that is forwarded to the storage is defined by the solar multiple and the control strategy. At design conditions the HRSG is supplying the power block required energy to generate the 4 MWe net and the excess is stored. The ceramic st for large contact surfaces so heat transfer is optimized. The air can flow through the storage in both directions (charging and discharging) but the storage temperature varies and discharging. Due to Ebsilon restrictions t he storage module was divided into a cold and a However in the power plant there is a single storage device with the profiles presented in Figure 2.31 [27] for charging and discharging Model of an atmospheric volumetric CRS Page 88 The quantity of energy that is forwarded to the storage is defined by the solar multiple and the control strategy. At design conditions the HRSG is supplying the power block required energy to generate the 4 MWe net and the excess is stored. The ceramic st orage is designed for large contact surfaces so heat transfer is optimized. The air can flow through the storage in both directions (charging and discharging) but the storage temperature varies during charging he storage module was divided into a cold and a However in the power plant there is a single storage device with the for charging and discharging .
2. Model of an atmospheric volumetric CRS Study of a hybrid concentrating solar power plant for Portuguese conditions Page 89 Figure 2.32: Ebsilon storage temperature profile. The storage output temperature is thus influenced by the HTF properties (temperature, flow rate), the actual available storage capacity and the storage design properties (e.g. materials, layout, etc.). These temperature differences can be estimated by the dashed line in Figure 2.32, which can be defined by the following equations [27]: • !r# = • +0.1×∆• × Ê' '6 (2.32) • M!r# =• “ −0.1×∆• ×Ë1− Ê' '6 Ì (2.33) ∆• =• r, −• !M, (2.34) The hot air then flows from the storage or receiver to the HRSG and its exhaust is sent back to the receiver, closing the power plant air cycle. There are several possible HRSG configurations. The simplest configuration of HRSG is composed by an economizer, evaporator and a superheater, which is illustrated in Figure 2.33. There are several connections to the steam cycle, e.g. a turbine extraction to the deareator or a feedwater extraction to be injected in the superheated steam to prevent sudden temperature increases. Several controllers are considered to maintain stable temperatures.
Study of a hybrid concentrating solar power plant for Portuguese conditions Figure 2. 33 2.3 Software integration The design and annual simulation of atmospheric air volumetric central receiver power plants is an iterative process and can be performed with several simulations tools. Some simulation tools such as NREL SAM, DLR GREENIUS or the ECOSTAR methodology use a simplified and intuitive approach, with very low calculation times, that allow studying different power plant configurations for an overall technical and economical perspective. On the other side, there are some complex models and tools that allow studying individual parts but not the overall plant, e.g. solar radiation, solar field, receiver performance, power block, costs, etc. These models need more calculation time and require detailed estimation of the variables to increase p For the present work, models with good accuracy and reasonable calculation times were created, so that CRS power plants could be optimized up to the detailed engineering and definitive cost evaluation stage. With this purpose, three tools were s Layout Calculation (HFLCAL) for solar field optimization, Ebsilon Professional for power circuit optimization, and Excel for software compilation and economic evaluation. 2. Model of an atmospheric volumetric CRS Study of a hybrid concentrating solar power plant for Portuguese conditions 33 : Ebsilon model for the receiver and storage. Software integration The design and annual simulation of atmospheric air volumetric central receiver power plants is an iterative process and can be performed with several simulations tools. Some simulation tools such as NREL SAM, DLR GREENIUS or the ECOSTAR methodology use a simplified and intuitive approach, with very low calculation times, that allow studying different power plant configurations for an overall technical and economical perspective. On the other side, there are some complex models and tools that allow studying , with very good precision, individual parts but not the overall plant, e.g. solar radiation, solar field, receiver performance, power block, costs, etc. These models need more calculation time and require detailed estimation of the variables to increase p recision. For the present work, models with good accuracy and reasonable calculation times were created, so that CRS power plants could be optimized up to the detailed engineering and definitive cost evaluation stage. With this purpose, three tools were s elected: Heliostat Field Layout Calculation (HFLCAL) for solar field optimization, Ebsilon Professional for power circuit optimization, and Excel for software compilation and economic evaluation. Model of an atmospheric volumetric CRS Page 90 The design and annual simulation of atmospheric air volumetric central receiver power plants is an iterative process and can be performed with several simulations tools. Some simulation tools such as NREL SAM, DLR GREENIUS or the ECOSTAR methodology use a simplified and intuitive approach, with very low calculation times, that allow studying different power plant configurations for an overall technical and economical perspective. On the other , with very good precision, individual parts but not the overall plant, e.g. solar radiation, solar field, receiver performance, power block, costs, etc. These models need more calculation time and require detailed For the present work, models with good accuracy and reasonable calculation times were created, so that CRS power plants could be optimized up to the detailed engineering and elected: Heliostat Field Layout Calculation (HFLCAL) for solar field optimization, Ebsilon Professional for power circuit
2. Model of an atmospheric volumetric CRS Study of a hybrid concentrating solar power plant for Portuguese conditions Page 91 Microsoft Excel was selected as compilation software because it is highly integrated with Ebsilon. Several models were built and optimized using Ebsilon Professional, while the solar field was designed and optimized using HFLCAL, and power plant economics were defined in Excel. For each time step Excel sends user defined input variables to Ebsilon, runs the Ebsilon model and gathers the results back to Excel. The integration of HFLCAL, Ebsilon and Excel in the model is presented in Figure 2.34. Figure 2.34: Compilation of tools used for the power plant design and annual simulation. 2.4 Economic model 2.4.1 Levelized electricity cost The power plant LEC was calculated according to the IEA Method [25]. To compare the different approaches, a combination of LEC plus sensitivity analysis was used to choose the best alternative and analyse the impact of different variables on the generated electricity cost. The LEC is dependent on the power plant capital investment (CAPEX), debt interests and insurance rates, annual operation and maintenance costs, annual fuel costs and generated annual net electricity, according to Equation 2.35 [25]. Annual simulation Power block and integration Solar field optimization η Receiver Η Solar field Q Receiver A Solar field T air p air ϕ air DNI T air P air Q Receiver Optimization EBSILON®Professional Techno-economical models Manufacturers Operating power plants References Solar field area Receiver characteristics Solar multiple Storage size Power block characteristics η Gross η net P electric
2. Model of an atmospheric volumetric CRS Study of a hybrid concentrating solar power plant for Portuguese conditions Page 92 zÍ· = ¯ Πׯ¡¤ÊÏ-Ç Ð&Ñ -Ç Î¦›‚ Ê œ›< (2.35) where C rf is the insurance and debt interest coefficient rate; CAPEX is the capital expenditure for the power plant; K O&M is the annual operation and maintenance costs; K fuel is the annual fuel costs and E net is the annual net electricity. Power plant C rf was considered to be 9.88 %, for a 30 year lifetime expectance. CSP power plants lifetime, for economic calculations, is usually considered between 20 and 40 years. Above this period power plant degradation may require to substitute key components such as the mirrors, receiver, etc., which are unaccounted in the regular maintenance costs. Annual operation and maintenance costs were considered for local conditions, using a percentage of CAPEX for fixed O&M costs, variable O&M costs and water usage costs (mainly heliostat cleaning, etc.) [25, 28]. 2.4.2 Cash flow analysis A more comprehensive economic model was built, calculating the cash flows and finding the investment return rate and period, based on SAM [28]. This project cash flow analysis is dependent on financial factors (fraction of own capital used, amortization structure, debt payment structure and methods used for future cash flow calculation) and power plant technical factors (power degradation, equipment performance and cost). The estimations considered for the cash flow analysis were: a 30 % own capital, a 20-year loan with an interest rate of 8 %, a 1 % annual insurance rate, a linear amortization for 20 years, current national profit taxes and a conservative 1 % of power degradation. The cash flow was done for the 30 years of power plant life time and the cash flow year is indicated by the letter n ranging from n = 0 (year zero) to n = 30 (the last year). The electricity generated for year one is not constant for the remaining years. The power plants suffer degradation of equipments, reducing their performance, usually generating less electricity and therefore lower operational income. The annual generated electricity and the operational income are approximated by the following equation: ÍÒÓJÔÒJ•Ò = ÍÒÓJÔÒJ•1×1−CÓ•%•FÔ-Ò•F } (2.36) ÍÒÓJÕ•ÖŠ=ÍÖ,FÔ,ÔFJ¿1ℎ×8%ÔÒF•Ô88 € "Ùr (2.37) The power plant operational expenses are: operation and maintenance (O&M) costs, fuel costs, insurance, financing costs and local taxes. For solar only power plants, the fuel cost
2. Model of an atmospheric volumetric CRS Study of a hybrid concentrating solar power plant for Portuguese conditions Page 93 is zero. The O&M costs are divided into the Fixed O&M costs and the variable O&M costs and are defined by: ÔH%Ú&à = G.-Ò•Ö,-.FË € Ì+ÔH%,-.FÛ,•Û•,ÔFJË € "Ù Ì×yJ.FÜ,•Û•,ÔFJ¿1 (2.38) Ž•Ô•ŒÖÚ&à = Ž•Ô•ŒÖ,-.FÛÖ,FÔ,ÔFJÓÒ•F%Ë € 6Ùr Ì׈ÒÒŠ•Ö-ŠFÛŠFÃ1ℎ (2.39) ˆÒÒŠ•ÖÛ-ÛFJF•H = G-ÛFJÕ•ÖŠ•FÔ-Ò€×z-,•ÖF•H% (2.40) The insurance costs were considered constant for the power plant life cycle, as in the case of the LEC analysis with a fixed rate of 1 % of the CAPEX [25]. To define the financing costs several concepts are important: Debt balance, Debt interest, Debt repayment and total payment. The debt balance for the first year is defined by: CŒFŒ•Ö•Ò, = −·ˆGÍÞ+EÒ,ÒFÔÕ.×CŒF8•,FÔ-Ò (2.41) If there are no incentives for the subsequent years the debt balance is defined by: CŒFŒ•Ö•Ò,,ŠÒFJ•= −CŒFŒ•Ö•Ò,ÛÕÔ-Š.J•−CŒFÛ•JÜÒFÛÕÔ-Š.J• (2.42) The total debt payment per year for the power plant is the amortization to the bank loan (debt repayment) and the debt interest payment to the bank for loaning the money (debt interest payment). •-F•ÖÛ•JÜÒF = CŒFÔÒF.F+CŒFÛ•JÜÒF (2.43) CŒFÔÒF.F = CŒFŒ•Ö•Ò,+z-•ÒÔÒF.F•F (2.44) The last factors to be accounted in the power plant cash flow analysis are the national taxes. In Portugal (in 2012) the following values apply: a 23 % sales taxes (VAT), the local municipally applies a 0.7 % property tax and the corporate profit taxes are 25 % or 12.5 %, depending on the company taxable income being above 12 500 € or not, respectively. The amortization of the equipment was considered with an annual tax of 5 %. This tax model is simplified and a more detailed model should be conducted in the future, as further tax savings can be obtained. Also the Portuguese tax model has been suffering some alterations in recent
3. Optimization of a 4 MWe atmospheric volumetric CRS power plant Study of a hybrid concentrating solar power plant for Portuguese conditions Page 100 This page was intentionally left blank
3. Optimization of a 4 MWe atmospheric volumetric CRS power plant Study of a hybrid concentrating solar power plant for Portuguese conditions Page 101 3. Optimization of a 4 MWe atmospheric volumetric CRS power plant The CRS optimization process involves the power block, design DNI, receiver flux, solar multiple, storage capacity and control strategy. Only a few options are presented in this thesis, as well as the results that allowed the selection of the best configuration decision for SOLMASS project. The start case for this study was the Jülich solar tower design scaled-up to the 4 MWe scale. The work presented in this Chapter (and the previous chapter), resulted into the manuscript [1]. 3.1 Options analysed 3.1.1 Optimization of power block, design DNI and receiver flux For the power block, design DNI and receiver flux optimization, several configurations and operating conditions were simulated. The single pressure HRSG systems analysed were: #1. a HRSG that generates 27 bar and 415 °C steam to feed a 2 stage turbine; #2. a HRSG that generates 60 bar and 450 °C steam to feed a 2 stage turbine; #3. a HRSG that generates 80 bar and 480 °C steam to feed a 3 stage turbine; #4. a HRSG that generates 130 bar and 530 °C steam to feed a 3 stage turbine; The multi-pressure HRSG systems analysed were: #5. a 2-pressure HRSG that generates 27 bar and 485 °C steam – based on the Jülich power plant [2] - to feed a 3 stage turbine; #6. a system based on the previous designed of PS10 project [3], considering a 2-pressure HRSG that generates 80 bar and 515 °C steam to feed a 4 stage turbine; #7. based on developments from Brightsource that announced their CRS with direct steam generation operated at 530 °C e 130 bar [4]. A similar steam condition 2-pressure HRSG with reheat was simulated to feed a 4 stage turbine at 530 °C and 130 bar. Options CRS#1 and CRS#5 were based on Jülich solar tower and were used to analyse the impact of multi-pressure HRSG in the technical and economic performance of CRS power plants. Options CRS#2, CRS#3 and #6 were designed to support new research vectors such as the integration of steam generated from the CRS into biomass power plants with similar conditions [5]. Forest waste biomass power plants with these steam characteristics are operating and projected in Portugal [6]. Options CRS#4 and CRS#7 correspond to high pressure and temperature systems, reaching the borderline established for open volumetric CRS and
3. Optimization of a 4 MWe atmospheric volumetric CRS power plant Study of a hybrid concentrating solar power plant for Portuguese conditions Page 102 commercial power blocks [3]. The objective is to operate with high pressure and temperature steam (to improve efficiency) and fully use all the potential of the high temperature air obtained from the open volumetric receiver. Due to water scarcity, a common scenario in high DNI locations (although not in Portugal), an air-cooled steam condenser was considered for all options. Table 3.1 presents the most important design variables considered, to analyse power block cycle selection impact. Table 3.1: Design values for the power block cycle selection impact analysis. Option CRS#1 CRS#2 CRS#3 CRS#4 CRS#5 CRS#6 CRS#7 Receiver Flow rate (ton/h) 115 104 108 110 116 97 95 Temperature (°C) 680 680 680 680 680 680 680 HRSG evaporator Pinch Point (°C) 17 17 17 17 31 (HP) 3 (LP) 31 (HP) 3 (LP) 31 (HP) 3 (LP)) Approach Temperature (°C) 22 22 22 22 39 (HP) 5 (LP) 39 (HP) 5 (LP) 39 (HP) 5 (LP) HRSG output Temperature (°C) 415 450 480 530 485 (HP) 400 (LP) 515 (HP) 400 (LP) 530 (HP) 400 (LP) Pressure (bar) 27 60 80 130 27 (HP) 15 (LP) 80 (HP) 40 (RH) 15 (LP) 130 (HP) 40 (RH) 15 (LP) Flow rate (ton/h) 23 20 20 20 20 (HP) 2 (LP) 15 (HP) 15 (RH) 3 (LP) 13 (HP) 13 (RH) 4 (LP) Condenser Temperature (°C) 49 49 46 46 46 46 46 Pressure (bar) 0.12 0.12 0.1 0.1 0.1 0.1 0.1
3. Optimization of a 4 MWe atmospheric volumetric CRS power plant Study of a hybrid concentrating solar power plant for Portuguese conditions Page 103 The solar field layout is influenced by several variables: equipments and space (e.g. limitations on the field available, heliostat beam quality), environmental conditions (e.g. available DNI, solar angles, temperatures, wind, humidity [7]) and by design definitions (e.g. design point DNI, heliostat distribution algorithm, heliostat/receiver dimensions, tower height and inclination). One of the variables is the design DNI. If the selected design DNI is high, energy dumping would be lower but the power plant would operate in partial load a larger period of time, with lower efficiencies. If the design DNI is low, dumping would be higher but the number of full power hours would also be larger. This can be confirmed in Figure 3.1. Figure 3.1: Cumulative hours per incident DNI for Faro, Portugal. With this perspective, and to obtain the optimal design DNI, four different DNI design values were tested for the CRS#3 power plant option: #8. design DNI 675 W/m 2 and receiver area of 72.6 m 2 ; #9. design DNI 825 W/m 2 and receiver area of 55.3 m 2 ; #10. design DNI 900 W/m 2 and receiver area of 50.8 m 2 . The heliostat dimensions and tower height were kept unchanged (respectively 60 m 2 and 97 m) and the remaining design definition variables were optimized to maintain similar receiver fluxes, average design fluxes of 550 kW/m 2 ; the resultant peak solar fluxes for option CRS#8 to CRS#10 were close to 950 kW/m 2 . As by experimental testing the receiver solar fluxes
3. Optimization of a 4 MWe atmospheric volumetric CRS power plant Study of a hybrid concentrating solar power plant for Portuguese conditions Page 104 could be higher (≈1000 kW/m 2 ) [8], to optimize the performance, a design case was considered with a 10 % solar flux tolerance in the receiver before the solar field defocus – option CRS#11: #11. design DNI 750 W/m 2 and receiver area of 60.0 m 2 with 10 % tolerance in the receiver peak flux. Option CRS#11 with a larger storage (3 hour at nominal load) was used to study the integration of different biomass based options into open volumetric CRS (option CRS#12), Chapter 4. 3.1.2 Optimization of solar multiple, storage capacity and control strategy Several options were analysed to find the best solar multiple, storage capacity and control strategy. For each option (set of solar multiple/ storage capacity/ control strategy), different receivers and blowers were simulated. The tested sets of solar multiple/ storage capacities were: • solar multiples of 1; 1.25; 1.5; 1.75; 2; • storage capacities of 1; 2; 3; 4; 5; 6 equivalent hours for each solar multiple. For each solar multiple, a different solar field configuration and layout was optimized using HFLCAL. As for higher solar multiples larger intercept power and solar field aperture are necessary, two alternatives were considered: a different distribution of the heliostats’ focal points in the receiver or, if it is insufficient, the receiver area was increased. For the design DNI of 750 W/m 2 the receiver dimensions obtained were: • solar multiple of 1 – receiver area of 48.6 m 2 ; • solar multiple of 1.25 – receiver area of 60.0 m 2 ; • solar multiple of 1.5 – receiver area of 72.6 m 2 ; • solar multiple of 1.75 – receiver area of 85.0 m 2 ; • solar multiple of 2 – receiver area of 96.8 m 2 . Solar multiple and storage capacity are important design variables in the CRS optimization process. Their optimization is also dependent on the operational strategy defined by the power plant responsible. For CRS power plants with large storage devices (or hybrid), a larger staff is necessary for 24 hours daily operation, while power plants with smaller storage devices need a smaller staff team. But it is only possible to find the best control strategy for a specific CRS via an optimization process, involving the thermal storage capacity, solar multiple,
3. Optimization of a 4 MWe atmospheric volumetric CRS power plant Study of a hybrid concentrating solar power plant for Portuguese conditions Page 105 energy dumping, feed-in tariff or the need to generate power on demand. Four different control strategies were considered. The first one is: • control strategy #1 (CS#1) – it is the most common control strategy in commercial CRS; it uses the solar power to run the power block and the excess heat is stored; this stored energy is used to cover solar transients and extend operation until storage is empty; the staff is scheduled for 2 shifts with extra-hours for extended operation. In CS#1, during power plant start-up, there is a period of time when solar radiation is not enough to run the power block and begin generating electricity (CS#1, nº1, Figure 3.2). This period is dependent on the inertia of components and on the available solar radiation. To minimize this start-up period, a different control strategy can be used taking advantage of residual energy stored from the previous day. Another alternative to reduce this start-up period and avoid transient problems is the integration of a fuel burner (full or partial hybridization). After start-up, the power block begins generating electricity, and, because the power plant is designed for solar multiples higher than 1, heat from the solar receiver surpasses the needs from the power block. During this period (CS#1, nº 3, Figure 3.2), the excess energy is stored while the power block is generating electricity at nominal power. When the storage device maximum capacity is reached, the excess energy is dumped, normally by defocusing heliostats from the receiver. In the evening (CS#1, nº5, Figure 3.2), the axial blowers reverse the flow and use the energy stored to compensate the solar radiation scarcity, extending power plant operation until the storage is empty. Figure 3.2: Application of control strategy CS#1 to a typical operating day. 4 hours Thermal Power (MW) 51 3 Q tS – Resultant thermal power to storage Q fS – Thermal power from storage Q D – Thermal storage dumping power 21 Plant down Partial load Full load – Charging storage Full load – Storage maximum reached - dumping Full load – Discharging storage 4 5 1 3 2 Q Rec – Thermal power from receiver Design Power
3. Optimization of a 4 MWe atmospheric volumetric CRS power plant Study of a hybrid concentrating solar power plant for Portuguese conditions Page 106 The decision diagram for CS#1 is based on C.J. Winter decision diagram [9], Figure 3.3. Figure 3.3: Decision diagram for CS#1 control strategy on a CRS.
3. Optimization of a 4 MWe atmospheric volumetric CRS power plant Study of a hybrid concentrating solar power plant for Portuguese conditions Page 107 Other control strategies considered are control strategy #2 (CS#2), with the power plant operating only in two fixed shifts (6:00 to 22:00) and the remaining energy being stored for the next day start-up, and control strategy #3 (CS#3) with the power plant operating only in one shift and reduced personnel costs: • control strategy #2 (CS#2) – power plant operates during 2 shifts without extra hours (6:00 to 22:00 hours), storing any excess heat for power plant start-up in the next day; • control strategy #3 (CS#3) – power plant operates during 1 shift (8:00 to 16:00 hours) plus 2 extra hour operation when necessary (16:00 to 18:00), storing any excess heat for power plant start-up in the next day; The aim of control strategies #1, #2 and #3 is to use the CRS for base load power, with the power plant generating an almost continuous electricity flow to the grid. However, 24- hour operation of an atmospheric air volumetric CRS is not common, due to the thermal storage cost and size. A different perspective for CRS power plants is to generate power on demand, mainly to support the peak electricity consumption periods with bonus feed-in tariffs. Figure 2.30 illustrated the electricity consumption in Portugal for a typical summer day, with consumption peaks from 9:00 to 13:00 and 18:00 to 21:00 (these peaks are even more pronounced during the winter). CSP can also be used to cope with this demand defining a control strategy for these conditions: • control strategy #4 (CS#4) – the power plant operates at nominal load (4 MW e ) only during hours with high network electricity demand (9:00 to 13:00 and 18:00 to 22:00) and at minimum power block load (2 MW e ) during the remaining period. 3.2 Results and discussion 3.2.1 Optimization of power block, design DNI and receiver flux There are no commercial open volumetric receiver CRS 4 MW e power plants in operation worldwide. The most important operating power plant with this technology is the 1.5 MW e Jülich Solar Tower [10]. In off-design conditions, the solar field performance was obtained by HFLCAL, and the receiver performance was approximated by a model based on experimental data from DLR (Chapter 2.1.4.4); the storage performance was approximated by a constant loss factor (Chapter 2.2.5.3); the HRSG pressure drops were obtained based on data from manufacturers and references; turbine performance was checked with manufacturers for option CRS#2 (2 stage steam turbine) and CRS#4 (3 stages steam turbine) (Chapter 3.4) [11].
3. Optimization of a 4 MWe atmospheric volumetric CRS power plant Study of a hybrid concentrating solar power plant for Portuguese conditions Page 108 Options CRS#1 and CRS#3 were considered to have similar operating nominal isentropic and mechanical efficiencies as options CRS#2 and CRS#4. For options CRS#1 to CRS#7 the receiver design dimensions were considered to be similar to the dimensions of a 60 m 2 heliostat, with design DNI of 750 W/m 2 , 1.25 solar multiple and 2 hour storage. Table 3.2 presents the main energy and economic results for CRS#1 to CRS#7 power blocks. Table 3.2: Power block operating conditions for a 4 MWe atmospheric air volumetric CRS. Option CRS#1 CRS#2 CRS#3 CRS#4 CRS#5 CRS#6 CRS#7 Performance Gross power (MW e ) 4.2 4.2 4.6 4.9 4.5 4.6 4.7 Net power (MW e ) 3.8 3.7 4.0 4.1 4.0 4.0 4.0 Annual net electricity generated (GWh) 10.3 11.1 11.6 12.1 10.8 12.4 12.6 Efficiency Power block cycle (gross) 21 % 23 % 24 % 26 % 22 % 26 % 27 % Power block cycle (net) 19 % 20 % 21 % 22 % 20 % 23 % 24 % Power block and CRS costs CAPEX PB cost (Million €) 2.8 2.9 3.1 3.7 3.7 4.3 4.4 CAPEX CRS cost (Million €) 22.0 22.1 22.3 23.0 23.0 23.8 23.9 Annual O&M cost O&M cost (thousands of €) 506 509 517 526 516 526 528 Table 3.2 indicates that, for single pressure HRSG, higher operation temperatures and pressures (options CRS#1 - CRS#4) resulted in improvements of power block cycle efficiency (up to 5 %). The increase in pressure also increases the parasitic losses, so the net efficiency gain is reduced up to 3 % (Table 3.2). However, even in single pressure HRSG, the utilization of higher pressure systems increases the system complexity, the equipments used are more expensive (increase in CAPEX, Table 3.2) and with higher annual maintenance costs (Table 3.2).
3. Optimization of a 4 MWe atmospheric volumetric CRS power plant Study of a hybrid concentrating solar power plant for Portuguese conditions Page 109 The power block (PB) impact in the power plant LEC is significant. Two methodologies were used to analyse the PB cost: a simplified PB cost, obtained by downsize of the ECOSTAR data to 4 MW e (Table 3.9), while keeping the relative cost constant for all operating pressures; and UPORTO model developed by the author (Table 3.7), Figure 3.4. Figure 3.4: LEC variation with power block operating conditions (lines are for readability). The same reasoning concerning the relation between cost and performance applies to multi-pressure systems, which generally exhibit higher efficiencies for higher pressures. Though, there is a significant overall increase in the cost of multi-pressure systems that justifies the need of a detailed model for the power block cycle cost. The LEC is lower for these multi-pressure systems; however, the LEC analysis does not consider equipments degradation and consequent performance decrease. The option with the best LEC is CRS#7 (0.229 €/kWh), Figure 3.4. Lower pressure and less complex power blocks, such as options CRS#1 and CRS#2, have a significantly higher LEC, 0.260 €/kWh and 0.243 €/kWh, respectively. Nevertheless, this complexity/cost trade-off should be a factor to analyse in the risk assessment done by the project investor. Between the single-pressure and multi-pressure HRSG there are significant changes in performance and in investment (CAPEX). Multi-pressure HRSG systems have higher costs and O&M than singlepressure HRSGs, but multi-pressure HRSGs are more energy efficient. Despite these differences, their LEC does not vary significantly, e.g. from option CRS#7 to CRS#4 there is a 0.003 €/kWh LEC difference. 0,200 0,210 0,220 0,230 0,240 0,250 0,260 0,270 1234567 LEC (€/kWh) Option LEC - € /kWh - UPORTO PB cost LEC - € /kWh - Simplified PB cost
3. Optimization of a 4 MWe atmospheric volumetric CRS power plant Study of a hybrid concentrating solar power plant for Portuguese conditions Page 116 Table 3.6: Power plant cash flow analysis and economic indicators. Plant designation Feed - in tariff (€/kWh) IRR (%) NPV (million €uro) * Payback period (years) * CRS#3_SM1.25_2S_CS#1 0.273 (24 hour) 9.8 % 7.9 14 CRS#3_SM1.5_4S_CS#1 0.273 (24 hour) 9.6 % 9.4 15 CRS#3_SM1.75_5S_CS#1 0.273 (24 hour) 9.3 % 10.2 16 CRS#3_SM2.0_7S_CS#1 0.273 (24 hour) 9.0 % 11.4 16 CRS#3_SM1.25_2S_CS#2 0.273 (24 hour) 9.8 % 7.8 15 CRS#3_SM1.25_2S_CS#2 0.341 (from 9 to 22 hours) 0.164 (remaining) 17.0 % 16.9 7 CRS#3_SM1.25_2S_CS#3 0.273 (24 hour) 7.6 % 4.9 20 CRS#3_SM1.25_3S_CS#4 0.273 (24 hour) 6.1 % 3.1 23 CRS#3_SM1.25 _3S_CS#4 0.341 (from 9 to 22 hours) 0.164 (remaining) 11.6 % 10.7 12 CRS#3_SM1.25 _3S_CS#4 0.341 (from 9 to 13 and 18 to 22 hours) 0.164 (remaining) 8.2 % 6.0 18 * - considering average inflation of 4 %. The investment in CRS power plants is attractive for the selected case (CRS#3_SM1.25_2S_CS#1, Table 3.6), with high IRR (9.8 %) and moderate payback time (14 years), and good NPV for the power plant life cycle (€ 7.9 million) even considering a conservative average inflation of 4 % (well above the December 2012 inflation - 2.2 %, 2.3 % for the Euro Area and European Union, respectively [12]). Power plants with better NPV (up to € 11.4 million) can be considered, but the CAPEX for these power plants is higher and the investment payback time is also higher.
3. Optimization of a 4 MWe atmospheric volumetric CRS power plant Study of a hybrid concentrating solar power plant for Portuguese conditions Page 117 In a different perspective, if the objective is to generate electricity adjusted to the network demand (larger generation during demand peak hours), the power plant would only be viable if a bonus feed-in tariff is obtained (CRS#3_SM1.25_3S_CS#4, Table 3.6). If the bonus tariff is obtained for the period between 9:00 to 13:00 and 18:00 to 22:00 hours, the investment IRR and NPV are below CS#1; however, if the bonus tariff is obtained for the period from 9:00 to 22:00, the investment IRR and NPV are significantly higher than CS#1. 3.3 Optimized 4 MWe CRS power plant The selected 4 MW e CRS configuration for Faro conditions is a CRS with design DNI of 750 W/m 2 , receiver dimensions of 60 m 2 , receiver peak flux of 950 kW/m 2 , SM of 1.25, 2 hours storage, a HRSG which generates 80 bar and 480 °C steam to feed a 3 stage turbine and using control strategy CS#1 – option CRS#3 or CRS#3_SM1.25_2S_CS#1. For this power plant a typical operational day is presented in Figure 3.9. Figure 3.9: Typical operational day for the 4 MWe CRS with 1.25 SM and 2 hours storage for CS#1 (lines are for readability). The power plant has an initial period when solar energy is available but the power plant is not generating electricity, followed by a period of partial load operation until the power plant begins operating at nominal load (at 7:00 – Figure 3.9). After this period of time, the power plant uses the excess energy to fill the storage (with 5 MWth from 8:00 to 14:00). After this period the storage capacity is full and the excess energy is dumped. At the end of the day, available solar energy is reduced and the storage flow is reversed, extending power plant -10 -5 0 5 10 15 20 25 30 35 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 Energy (MW) Hour QInt QRec QPB Egross Enet Qs
3. Optimization of a 4 MWe atmospheric volumetric CRS power plant Study of a hybrid concentrating solar power plant for Portuguese conditions Page 118 operation until 20:00 (Figure 3.9). The 1.25 SM and 2 hours storage 4 MW e CRS annual performance using CS#1 (CRS#3_SM1.25_2S_CS#1) is presented in Figure 3.10. Figure 3.10: 4 MWe CRS with 1.25 SM, 2 hours storage and CS#1 annual efficiency. The solar field annual efficiency (63 % - Figure 3.10) includes the heliostat stow positioning losses (caused e.g. by excessive wind) and the losses due to the solar DNI upper limit. The receiver annual efficiency is 77 % - Figure 3.10 - and the storage unit has an annual efficiency of 85 % - Figure 3.10. The storage unit efficiency includes the dumping losses when excess energy is available but the storage capacity is fulfilled. Up to the power block input the accumulated energy efficiency is 41 % (Figure 3.10). The component with the lowest efficiency is the power block (29 % gross - Figure 3.10); its efficiency could be improved using a combined cycle (gas turbine plus steam turbine) but it would imply changing the receiver technology (e.g. pressurized air receiver) or to consider hybrid solutions. The scale-up of the power plant could also increase power block efficiency, as more efficient turbines can be used. The accumulated efficiency, solar to electricity, is 12 % (Figure 3.10). The parasitic loss, with the consumption of electric equipment, was 1.6 GWh per year, with significant contributions of the air-cooled condenser and blowers. If these parasitic losses are taken into account, the overall solar to electricity efficiency is 10 % (Figure 3.10). Larger power plants can improve the solar to electricity efficiency up to 20 % [13]. The CAPEX for the optimized 4 MWe CRS power plant is expected to be 22.3 M€, with the component costs presented in Table 3.7.
3. Optimization of a 4 MWe atmospheric volumetric CRS power plant Study of a hybrid concentrating solar power plant for Portuguese conditions Page 119 Table 3.7: CRS component costs – option CRS#3_SM1.25_2S_CS#1. The main components cost share is presented in Figure 3.11. Figure 3.11: Cost structure of a 4 MWe atmospheric air volumetric CRS - option #3. The main costs associated with the power plant are the solar field (36 %), the solar receiver (16 %), the power block (14 %) and storage (10 %). This indicates the importance of optimizing the power block cycle, receiver design DNI and flux, storage capacity, solar multiple and control strategy. Direct costs (DC) UPorto model - 4 MW unit at Faro/Tavira – for best option Solar field 150 €/m 2 [14] Storage 64 €/kWh th Receiver 120 €/kW th - based on receiver power and dimensions) [15] Tower 1 000 000 € [16] Power block 671 €/kW e -based on components design/operating conditions [17,18] Land 3.6 €/m 2 [16] Indirect costs (IC) Local work costs – 20 % of DC
3. Optimization of a 4 MWe atmospheric volumetric CRS power plant Study of a hybrid concentrating solar power plant for Portuguese conditions Page 120 3.4 Validity The power plant CAPEX is divided into the direct costs (DC) and indirect costs (IC). Direct costs are the cost attributed to equipment acquisition and their installation. It can be divided into several groups of components: heliostat field, power block, storage, solar receiver, tower and land. Indirect costs are the costs that are not directly attributed to a specific object, as the case of the start-up costs and the surcharge for construction, engineering & contingencies. For the atmospheric volumetric CRS one of the most complete cost reports is the ECOSTAR report [3], Table 3.8. Table 3.8: ECOSTAR reference costs for an atmospheric volumetric CRS power plant. EC OSTAR - 10 MW unit at Seville [3] ECOSTAR - 50 MW unit at Seville [3] Direct costs (DC) Solar field 150 €/m 2 138 €/m 2 Storage 60 €/kWh th 54 €/kWh th Receiver 115 €/kW th 103 €/kW th Tower 2 000 000 € 8 934 538 € Power block 600 €/kW e 536 €/kW e Land 2 2 €/m 2 Indirect costs (IC) Flat rate 20 % of DC 20 % of DC Life cycle 30 years 30 years Debt interest 8 % 8 % Insurance 1 % 1 %
3. Optimization of a 4 MWe atmospheric volumetric CRS power plant Study of a hybrid concentrating solar power plant for Portuguese conditions Page 121 The values presented in Table 3.8 are for a power plant located in Seville, Spain with a SM=1.82, and based on PS10 design done by KAM to SOLUCAR, with commercial equipment price consultations. The data presented in Table 3.8 cannot be directly used for the SOLMASS CRS, because the installed power is only 4 MWe. So, the component costs have to be extrapolated from the ECOSTAR data for the 4 MWe case. They can be estimated adjusting a cost curve, with scale factors based on the ECOSTAR report [3] (solar field – 0.95; receiver – 0.87; storage - 0.93). · Ê,à = · Ê,Ù Ë Ï á Ï > Ì ∝ (3.1) where C E,Y is the cost of equipment at the required size or capacity, C E,W is the cost of the reference equipment at reference size or capacity, X Y is the size or capacity of the required equipment, X W is the size or capacity of the reference equipment and α is the scale factor. Using the data from Table 3.8, the equipment costs for a 4 MW power plant were calculated and are presented in Table 3.9. Table 3.9: Economic data for the CRS model. Direct costs (DC) Simplified model - 4 MW unit downsize (using Equation 3. 1) Solar field 155 €/m 2 Storage 64 €/kWh th Receiver 119 €/kW th Tower 890 572 € Power block 638 €/kW e Land 2 €/m 2 Indirect costs (IC) Flat rate – 20 % of DC As different configurations were analysed, detailed models were developed to estimate the cost of the equipments. The solar field cost was calculated according to the Helios3s design from UPorto [14], Table 3.7. Heliostat costs are within the range presented at the SolarPaces heliostat catalogue [19] and according to the costs used for a 10 MW e power plant in ECOSTAR report [3]. A downsize to 4 MW e based on ECOSTAR report scale factors,
3. Optimization of a 4 MWe atmospheric volumetric CRS power plant Study of a hybrid concentrating solar power plant for Portuguese conditions Page 122 results in a 5 €/m 2 higher cost than Helios3s, but, considering the contacts done with EPC contractors and heliostat developers, the cost of 150 €/m 2 was pointed as realistic for the current commercial heliostat market [20]. Commercial heliostats, e.g. Colon 70 and Sanlucar 90, are operating at Plataforma Solar Almeria (PSA) and had a cost of 130 USD/m 2 (year 2000), for a 1000 heliostat annual production line. Assuming CE index factors of 394 for 2000 and 585.7 for 2011, with an average exchange rate USD/EUR of 0.72, the actual heliostat cost should be near 139 €/m 2 , which is 11 % lower than the cost considered. The storage device cost was based on the ECOSTAR report [3]. Novel storage with ceramics can reduce the cost below 20 €/kWh th [21], and so the storage cost used was considered as very conservative. ECOSTAR receiver cost model is based on the cost per incident power; as different design DNIs and receiver dimensions were selected for the same operating power, a detailed model was created, based on the receiver dimensions and incident power, UPORTO model. The simplified model used cost data from ECOSTAR [3] and a scale factor for the power/cost relation (Equation 3.1). The UPORTO model used data from Solair and Jülich projects [15] to estimate a cost per receiver components (cup, structure, insulation, and tubing) and to find a scale factor for the receiver area and power/cost. Solving Equation 3.1 for each receiver option, led to cost factors (€/m 2 and €/kWth) for the receiver and its respective estimated cost. The land and tower costs were checked by a national company [16]. Because different power block configurations were studied a detailed power block economic model was developed, based on the individual factors from the Guthrie’s method [17, 18], using Ebsilon detailed mass and energy balance simulations for equipment sizing, definition of construction materials and labour costs – UPORTO model. This method is referred by Peters et al. [18] as a definitive estimate with a good accuracy (error ± 10 %) to start detailed engineering and definitive cost evaluation. The indirect cost model is based on Guthrie’s method [17, 18]. The model used is composed by several components and is influenced by several factors: capacity, material and operational, Table 3.10.
3. Optimization of a 4 MWe atmospheric volumetric CRS power plant Study of a hybrid concentrating solar power plant for Portuguese conditions Page 123 Table 3.10: Power block components cost - for best option. Equipment Power factor Material factor Operational factor Cost (×10 3 USD 2000 ) Condensate economizer 1.2 MW Steel 1.5 bar (condensate) 20 Economizer (HRSG) 3 MW Steel 1.5 bar (steam) 91 Evaporator (HRSG) 9 MW Steel 80 bar (steam) 219 Super-heater (HRSG) 2 MW Steel 80 bar (steam) 16 Super-heater (HRSG) 1 MW Steel/ Cr -Mo 80 bar (steam) 21 Condenser 12 MW Steel Air cooled condenser 110 Turbine 4.6 MW - - 716 Generator 4.6 MW - - 9 Deareator - - - 19 Pumps 2 units - - 48 Drives for pumps 2 units - - 4 Fans 2 units - - 33 TOTAL 1 307 The values are in US Dollar ($) for 2000 with a CE cost index of 394. An actualization of the values to 2011 is necessary, using a cost index of 586 [22]; it is also necessary to convert to EUR using the 2011 average exchange rate of 0.72 EUR for each USD. The component costs in Table 3.10 do not include the costs associated with transport, insurance, assembly, cables, insulation, etc. The method used considers these costs in the bare module factor, which is multiplied by the equipment acquisition costs, Table 3.11 [17, 18].
3. Optimization of a 4 MWe atmospheric volumetric CRS power plant Study of a hybrid concentrating solar power plant for Portuguese conditions Page 124 Table 3.11: Bare module factor for power block equipments. Equipment Bare module factor Fans 2.19 Boilers Pre - fabricated 2.19 Locally assembled 1.86 Heat exchangers 3.17 Evaporators 2.45 Air cooled condenser 2.2 Pumps 3.3 Electric generators 2.5 Turbines: steam and gas 3.5 Pressure vessels Vertical 4.16 Horizontal 3.05 This method is used to get an estimate of the components costs. The final equipment cost is usually done via a public consultation with manufacturers and suppliers. Because several possibilities were analysed, the reference costs of the main components (HRSG and steam turbine) were requested to some manufacturers for several design cases. The approach used to validate the economic model with real data was to adjust the bare module factor so the equipment cost obtained via model matches the commercial quotation. The validation of the performance of the different power block options analysed was also done via consultation with manufacturers and reliable references. The power block used on Jülich solar tower uses a HRSG that generates 485 °C and 27 bar steam to feed a steam turbine/generator of 1.5 MWe. As at the time the Jülich solar tower was on early commercial exploration phase, it was not possible to obtain the measured HRSG and turbine characteristic curves. Several contacts were done with the power plant operators (KAM), but it was not possible to obtain this information because of confidentiality issues. HRSG manufacturers were contacted to obtain a solution that would fit the required specifications. The Rentech HRSG pre-assembled commercial series allows obtaining the pressure and temperature used in Jülich but with a steam flow rate of 10 tons/hour. For the 4 MWe power plant (option CRS#1) it would be necessary to have two equipments working in parallel, or a larger HRSG (typically operates at higher pressures and temperatures), or a specially designed HRSG. Option CRS#2 validated steam turbine was used for option CRS#1 (adjusting the operational conditions), and the necessary flow rate to obtain the design power was calculated. Option CRS#2 uses a
3. Optimization of a 4 MWe atmospheric volumetric CRS power plant Study of a hybrid concentrating solar power plant for Portuguese conditions Page 125 commercial SIEMENS SST-110 (2 stages) steam turbine, validated, according to the manufacturer, for the conditions reported on Table 3.12 [11]. Table 3.12: Validation data for the SIEMENS SST-110 turbine. Model Manufacturer data Input Temperature 450 °C 450 °C Pressure 60 bar 60 bar Flow 20.0 t/h 20.0 t/h Stage 1 Output temperature 196 °C 196 °C Output pressure 3.9 bar 3.9 bar Output flow 20.0 t/h 20.0 t/h Stage 2 Output temperature 49.4 °C 49.5 °C Output pressure 0.12 bar 0.12 bar Output flow 18.5 t/h 18.5 t/h Generator Frequency 50 Hz 50 Hz Power 4235 kW 4235 kW The partial load characteristic curves were requested but, as it is confidential information, only several performance values were given at certain loads [11]. These values were used and interpolated by Ebsilon characteristic lines to obtain the turbine performance at partial load, Figure 3.12. Figure 3.12: Performance characteristics of the steam turbine.