Long-term projections of electricity generation costs in Portugal

FACULDADE DE ENGENHARIA DA UNIVERSIDADE DO PORTO Long term Projections of Electricity Generation Costs in Portugal André Felício Duarte Beirão Carapito Mestrado Integrado em Engenharia Eletrotécnica e de Computadores Supervisor: Professor Doutor Cláudio Monteiro June 29, 2015 © André Felício Carapito, 2015 Abstract Due to the current aims of reducing CO2emission levels and increasing energy efficiency, the electric energy market is one of the principal sectors to approach big changes in the next years. In fact, the emphasis on renewable energy sources seems to be the most adequate response to this challenge; however doubts exist regarding the sustainability and economic viability of these solutions. Within this panorama, the present dissertation deals with the study of future electricity production costs for each type of production technology. Having these projections and the Portuguese electric system´s historic background in mind, this research aims at foreseeing the electricity production costs in Portugal, through the creation and adoption of various penetration scenarios of each type of technology in the national energetic mix. Here, the final aim is to understand the relation between these results and the additional costs that the system entails with the incentives paid to electricity producers. For that purpose, the methodology applied is based on the calculation of the levelized cost of electricity (LCOE) which considers the total costs, as well as the energy produced in each type of electricity generation technology from 2000 until 2030. The results of this investigation enabled the verification of a decreasing trend of the electricity production costs in case of renewable production technology sources, with a special focus on wind onshore and solar PV. Contrarily conventional thermal centrals demonstrate relatively higher costs caused by the increase in fuel and CO2emissions licenses prices. Furthermore, regarding system costs with electricity production, it was verified that they should persist at current levels, while in scenarios of high renewables penetration costs tend to diminish. In conclusion, all these findings confirm that renewable energy sources are a good solution in order to be able to fulfill the above-mentioned aims, while preventing an increase in prices for the system if the incentives for its development are applied in a sustainable manner. i ii Resumo Com as metas atuais de redução dos níveis de emissão de CO2e do aumento de eficiência energética, o mercado de energia elétrica é um dos principais setores ao qual se perspetivam fortes mudanças nos próximos anos. A aposta nas energias renováveis parece ser a resposta adequada ao problema mas surgem dúvidas quanto à sustentabilidade e viabilidade económica destas soluções. Enquadrada neste panorama a presente dissertação propõe-se a estudar o futuro dos custos de produção de eletricidade para cada tipo de tecnologia de produção. Com estas projeções e tendo em conta o historial do sistema elétrico português, pretende-se prever, através da criação de cenários de penetração de cada tipo de tecnologia no mix energético nacional, os custos de produção de eletricidade em Portugal. Como objetivo final procura-se perceber a relação entre estes resultados e os custos adicionais que o sistema acarreta com os incentivos pagos aos produtores de eletricidade. A metodologia seguida baseou-se no cálculo do custo nivelado de energia elétrica, ou levelized cost of electricity (LCOE), que tem em consideração os custos totais e a energia produzida em cada tipo de tecnologia de produção de eletricidade, desde o ano 2000 até ao ano 2030. Os resultados do estudo realizado nesta dissertação permitiram verificar que existe uma tendência decrescente dos custos de produção de eletricidade nas tecnologias de produção a partir de fontes renováveis, com especial foco na eólica onshore e na solar fotovoltaica. Contrariamente, as centrais térmicas convencionais deverão ter um custo mais elevado, proveniente do crescimento dos preços de combustível e de licenças de emissão de CO2. Relativamente aos custos do sistema com a produção de eletricidade verificou-se que no contexto atual estes deverão manter-se nos níveis atuais, sendo que em certos cenários de elevada penetração de renováveis os custos tendem a decrescer. Em suma, com estes resultados verifica-se que a aposta nas energias renováveis é uma boa solução para o cumprimento das metas anteriormente referidas, solução esta que não encarece o sistema desde que os incentivos ao seu desenvolvimento sejam realizados de forma sustentável. iii iv Acknowledgments A special thanks goes to the authors´ supervisor, PhD. Cláudio Domingos Monteiro for his total availability and guidance,allowing the completion of this thesis. To my friends and colleagues João Fernandes and Tiago Lima, who helped me to overcome this final phase. To all my family, specially my parents, for the continuous support they have given me and to my brother to whom I dedicate this thesis. Lastly, a very special thanks to Christiane Staender, for her unwearying support and motivation, which gave me strength to continue in the hardest moments. André Felício Carapito v vi LIST OF FIGURES xiii 6.4 Energy Produced by Solar PV Centrals in the "PV More" Scenario and in the RemainingScenarios................................. 67 6.5 Energy Produced by Onshore Wind Centrals in the "Wind More" Scenario and in theRemainingScenarios............................... 68 6.6 Energy Produced by Coal utilities in the Base Scenario, according to the case and methodconsidered.................................. 68 6.7 Energy Produced by NG utilities in the Base Scenario, according to the Case and MethodConsidered.................................. 69 6.8 Energy Produced by Coal and NG utilities in the "CHP remains" Scenario, according to the Case and Method Considered. . . . . . . . . . . . . . . . . . . . . . . 70 6.9 Energy Produced by Coal and NG utilities in the "CHP as efficiency" Scenario, according to the Case and Method Considered. . . . . . . . . . . . . . . . . . . 70 6.10 Energy Produced by Coal and NG utilities in the Solar PV Scenarios, according to the Case and Method Considered. . . . . . . . . . . . . . . . . . . . . . . . . . 71 6.11 Energy Produced by Coal and NG utilities in the "Wind More" Scenario, according to the Case and Method Considered. . . . . . . . . . . . . . . . . . . . . . . . . 71 6.12 Overall LCOE of each Electricity Generation Technology and System LCOE over thePastYears..................................... 73 6.13 LCOE Projections of all Electricity Generation Technologies for the Year of Installation. ...................................... 74 6.14 LCOE Projections of Coal Utilities in all the Created Scenarios. . . . . . . . . . 75 6.15 LCOE Projections of NG Utilities in all the Created Scenarios. . . . . . . . . . . 75 6.16 Overall LCOE Projections of RES and CHP. . . . . . . . . . . . . . . . . . . . 76 6.17 Overall LCOE Projections of Coal Utilities in all the Created Scenarios. . . . . . 76 6.18 Overall LCOE Projections of NG Utilities in all the Created Scenarios. . . . . . 77 6.19 System LCOE Projections in the Actual Panorama. . . . . . . . . . . . . . . . . 77 6.20 System LCOE Projections in the Base Scenario. . . . . . . . . . . . . . . . . . . 78 6.21 System LCOE Projections in the "CHP remains" and in the "CHP as efficiency" Scenarios....................................... 78 6.22 System LCOE Projections in the "PV More" and in the "PV as efficiency" Scenarios. 79 6.23 System LCOE Projections in the "Wind More" Scenario. . . . . . . . . . . . . . 79 6.24 System LCOE Projections in the Best Case of each Scenario. . . . . . . . . . . . 80 6.25 Projection of the Portuguese Electricity Spot Market Prices until 2030. . . . . . . 81 6.26 Overall LCOE and Total Additional Cost of Solar PV. . . . . . . . . . . . . . . . 81 6.27 Overall LCOE and Total Additional Cost of Wind Onshore. . . . . . . . . . . . . 81 6.28 Overall LCOE and Total Additional Cost of SHP. . . . . . . . . . . . . . . . . . 82 6.29 Overall LCOE and Total Additional Cost of CHP. . . . . . . . . . . . . . . . . . 82 6.30 Overall LCOE and Total Additional Cost of Biomass. . . . . . . . . . . . . . . . 82 6.31 Overall LCOE of Large Hydropower Plants and Total Additional Cost with CMEC. 83 6.32 Overall LCOE of Coal Utilities and Total Additional Cost with CAE. . . . . . . . 83 6.33 Overall LCOE of NG Utilities and Total Additional Cost with CAE. . . . . . . . 83 6.34 System LCOE (Base Scenario, Case A) and Total Additional Cost of the Electric System. ....................................... 84 A.1 Electricity Production Mix in Portugal in 2010. . . . . . . . . . . . . . . . . . . 89 A.2 Electricity Production Mix in Portugal in the Base Scenario (Case A). . . . . . . 90 A.3 Electricity Production Mix in Portugal in the "CHP remains" Scenario (Case A). . 90 A.4 Electricity Production Mix in Portugal in the "CHP as efficiency" Scenario (Case A)........................................... 91 xiv LIST OF FIGURES A.5 Electricity Production Mix in Portugal in the "PV More" Scenario (Case A). . . . 91 A.6 Electricity Production Mix in Portugal in the "PV as efficiency" Scenario (Case A). 92 A.7 Electricity Production Mix in Portugal in the "Wind More" Scenario (Case A). . . 92 A.8 LCOE of Coal Utilities Separated by each Component. . . . . . . . . . . . . . . 93 A.9 LCOE of Natural Gas Utilities Separated by each Component. . . . . . . . . . . 93 A.10 LCOE of CHP Power Plants Separated by each Component. . . . . . . . . . . . 93 A.11 LCOE of Large Hydropower Plants Utilities Separated by each Component. . . . 94 A.12 LCOE of SHP Plants Separated by each Component. . . . . . . . . . . . . . . . 94 A.13 LCOE of Geothermal Power Plants Separated by each Component. . . . . . . . . 94 A.14 LCOE of Wind Onshore Power Plants Separated by each Component. . . . . . . 95 A.15 LCOE of Solar PV Power Plants Separated by each Component. . . . . . . . . . 95 A.16 LCOE of Biomass Power Plants Separated by each Component. . . . . . . . . . 95 List of Tables 2.1 Characteristics of Top-down and Bottom-up Models.Source: Taken from [10]. . . 8 2.2 Some Examples of Energy Models. . . . . . . . . . . . . . . . . . . . . . . . . . 11 2.3 Main Electricity Markets Models and their Usage. Source: Adapted from [11]. . 13 2.4 Examples of LCOE Equation. . . . . . . . . . . . . . . . . . . . . . . . . . . . 17 2.5 Costs of New Electric Power Plant with and without CO2Capture on Current Technology (2007USD/EUR 0,73). Source: Adapted from [12]. ......... 24 2.6 Renewable Energy Support Policies. . . . . . . . . . . . . . . . . . . . . . . . . 26 2.7 Renewable Energy Support Policies by Country( EU-28, Norway and Switzerland)1Source: Adapted from [13]. ......................... 29 3.1 Utilities Lifetime. Sources: [3,6,7,14,15]...................... 38 3.2 Historical Prices for Natural Gas and Coal since 2000. Source: [16]. ....... 38 3.3 Forecast Prices for Natural Gas and Coal until 2030. Prices of 2011. Source: [17]. 39 3.4 Historical Exchange Rates, USD to EUR. Source: [18]............... 39 3.5 Natural Gas and Coal Prices in e/MWh from 2000 until 2030. . . . . . . . . . . 41 3.6 CO2Emission Rights (2011 Prices).Source : [17].................. 41 3.7 CO2Costs for Natural gas and Coal. . . . . . . . . . . . . . . . . . . . . . . . . 42 3.8 Annual Electricity Spot Market Prices and Additional Costs of PRE Technologies, CAE, CMEC and "Power Capacity Payment". . . . . . . . . . . . . . . . . . . . 44 4.1 Capacity factors of RES(excluding large hydropower). . . . . . . . . . . . . . . 46 A.1 Characteristics of past and future Hydropower plants in Portugal. . . . . . . . . . 87 A.2 Installed Capacities of SH in Portugal since 2000. . . . . . . . . . . . . . . . . . 88 A.3 Energy Produced by each Electricity Generation Technology in Portugal from 2000to2014inGWh. ............................... 89 A.4 Capital Costs in e/kW of the Different Electricity Generation Technologies Relevant for this Study. Source: [19–21]......................... 96 A.5 Fixed O&M Costs in e/kW of the Different Electricity Generation Technologies Relevant for this Study. Source: [19]......................... 97 A.6 Variable O&M Costs in e/kW of the Different Electricity Generation Technologies Relevant for this Study.Source: [19]. ..................... 98 xv xvi LIST OF TABLES Acronyms Btu British thermal unit CAPEX Capital expenditure CCGT Combined-Cycle gas turbine COE Cost of energy CHP Combined heat and power CSP Concentrating solar power EUR Euros GDP Gross domestic product GJ Gigajoule GW Gigawatt GWh Gigawatt per hour IEA International Energy Agency IGCC Integrated gasification combined cycle IRR Internal rate of return kW Kilowatt LCOE Levelized cost of electricity LHP Large hydropower MW Megawatt MWh Megawatt per hour MSW Municipal solid waste NEA Nuclear Energy Agency NG Natural gas OECD Organisation for Economic Co-operation and Development O&M Operation and maintenance OPEX Operational expenditure PC Pulverised coal PV Photovoltaic PRE Special regime production REE Spanish electricity network operator REN Portuguese electricity network operator RES Renewable energy sources SHP Small hydropower USD U.S. dollar ($) WWW World Wide Web xvii Chapter 1 Introduction 1.1 Context, Pertinence and Motivation After the liberalisation of the energy sector that opened its doors to a free and more competitive market of buying and selling electric energy in Portugal and most other countries around the globe, the very structure of the energy sector was no longer centralised in one company only, but in diverse firms. In the former structure, the planning has been made by considering the uncertainty of demand and the fuel prices only. Nowadays however, companies in the new market are increasingly confronted with new types of uncertainty and depend on the customers´ choices, as well as their competitors´ actions. Never before have the projection of the evolution of production costs, as well as electricity costs assumed that much importance as they assume today. Figure 1.1: Global Energy Issues in 2015.Source: [1]. 1 2Introduction Figure 1.1 presents a graphic that includes a variety of problems and their degree of uncertainty, as well as its impacts. Here, the upper corner on the right hand side distinguishes the energy prices. In fact, it is the pertinence of this issue that is quite a motivating factor regarding the realisation of this research. The commitment of a variety of nations around the world (such as Portugal) to reduce CO2 emissions in electricity generation, as well as to improve energy efficiency leads to a strong incentive to increase the production via Renewable Energy Sources (RES) and causes an energy mix that is quite different from the former one. Having this panorama in mind, this dissertation studies the evolution of electricity generation costs in Portugal until 2030. Furthermore, it is intended to provide answers to the following questions: • What is the future evolution of the electricity generation costs of each electricity production technology and which are the most competitive ones; • Knowing the background of Portugal, what is the future evolution of the electricity generation costs in this country in different scenarios; • And finally, once having obtained answers and results to the above-mentioned questions it is intended to know which implications the latter have on the additional costs of the electric system when deriving from incentives made to electricity producers. 1.2 Structure In order to be able to answer the above-mentioned questions and thus achieve the goals set for this research, this dissertation is structured as follows: In Chapter 1(the present one), an introduction to the dissertation topic, as well as the motivation and goals of the study get explained. Afterwards, in Chapter 2a review of the sate-of-the-art is conducted to help to internalise the several subjects important to the dissertation; Chapter 3subsequently presents relevant information regarding electricity production, consumption and international exchanges, as well as power plants costs and other characteristics, which has been collected for the conduction of this investigation. Chapter 4describes the adopted methodology for the calculation of the electricity generation costs of each electricity generation technology and evolution of the latter in Portugal until 2030. Chapter 5explains the several scenarios created, regarding future projections of installed capacity in Portugal, starting with a base scenario and how further scenarios have been developed on the basis of the first one. Hereupon Chapter 6demonstrates the results obtained for the projections of the electricity generation costs of each power generation technology, as well as the electricity generation costs in Portugal. Moreover it studies the implications of the results on the additional costs of the system. 1.2 Structure 3 Finally Chapter 7, presents the main conclusions, as well as the limitations of this research and recommendations for future investigations. 4Introduction 2.2 Long term Electricity Prices Forecasting: Methodologies. 11 Table 2.2: Some Examples of Energy Models. Models Features LEAP Long-range alternative energy planning (LEAP) is a tool that can be used to create models to perform a variety of tasks. It includes energy polices analysis, integrated energy planning, production of energy master plans, energy forecasting and energy scenario studies. It uses top-down strategy for demand and bottom-up for supply and follows an econometric methodology for the previous and simulation methodology for the latter. It covers local to global geography and is used for a medium or long-term scope. MARKAL/TIMES The model focuses only on the energy sector using a bottom-up strategy. It has two approaches to modelling energy: a technical engineering approach and an economic approach. Linear programming is used as mathematical approach over a medium to long-term time horizon. His purpose is to explore the possible energy futures based on several scenarios using an optimisation methodology. EnergyPLAN An deterministic model that optimises the operation of a given energy system on the basis of inputs and outputs defined by the user [28]. It is used to assist national or regional strategies plans in a long-term time horizon. Inputs of the model can be demands, renewable energy sources, capacities, costs and regulation strategies. As for outputs, usually are annual production values, energy balances, fuel consumption and system costs. IKARUS This model is a dynamical bottom-up linear cost-optimisation scenario for national energy systems. It is a long term study and can be applied for studying the effects of stochastic energy prices on long-term energy scenarios, carbon capture and storage in reducing carbon emissions, the introduction of fuzzy constraints to provide a better representation of political decision processes in the energy economy and energy policy and the implications of high energy prices. ORCED The purpose of this model is to dispatch power plants in a certain region to meet the electricity demands for any given year up to 2030 [29]. 2.2 Long term Electricity Prices Forecasting: Methodologies. Since countries began with the deregulation of the electricity generation market allowing several power companies to compete with each other, the importance of price forecasting has increased. Before, one central company performed a unique optimised plan for capacity expansion regarding the amount of new capacity, the generation mix and the timing of investments on a regional or national level [30]. This planning was only affected by uncertainty on future demand and fuel prices. Nowadays, however, companies are confronted with new uncertainties and depend on customer’s preferences and the action of competitors [31]. Therefore long term planning faces bigger challenges today. Electricity prices forecasting are divided by time horizon of the forecast into short-term, medium-term and long-term forecasting. There are no specific scopes for these horizons but usually short-term involves forecast from a few minutes up to a few days ahead, medium-term from a few days to a few months and long-term with a scope of months, quarters or years [32]. Literature on electricity prices forecasting is extensive, but most of the articles focus on short or medium- 12 State-of-the-art term forecast. However there are some main articles that refer which methodologies exist for the long-term horizon and others that use models to study the dynamics of electricity markets. The contents of these articles are in this section summarised focusing on information for the long-term electricity prices forecasting methodologies. The formulation of the electricity prices forecasting problem found in the reviewed articles differ in the terms applied but the approaches are similar. In Figure 2.2 it is presented a classification based on those made by Articles [2,11,33,34]. The different models dispose of diverse input data and results [2]. Figure 2.2: Electricity Price Forecasting Methodologies. A definition based on previously-mentioned articles articles for the several methodologies is given below: • Equilibrium analysis: These models use game theory to model the strategies of the market participants and how to solve them. The outcome of the study is the evolution of prices and the mathematical solution of the "games". The inputs and outputs of this model are visible in Figure 2.3. Examples of this type of model are the Cournot model, Bertrand model, and supply function equilibrium model; Figure 2.3: Equilibrium Analysis Model. Source: Taken from [2]. • Statistical methods: Time series data are the input and output of these models which use a statistical point of view to analyse prices evolution and do not take the physical processed 2.2 Long term Electricity Prices Forecasting: Methodologies. 13 into detailed account. The articles reviewed present a more specific classification of these methods which contains among others stochastic models, artificial intelligence based models, regression or causal models and econometric models. Statistical methods are commonly used for the hourly market clearing prices in a day-ahead spot market; • Simulation methods: When an equilibrium model cannot solve a problem because of its complexity, generally simulation methods are applied. The actual operation of the power system is simulated involving economical dispatch, computational power flow and physical system constraints. Most of the simulation methods are production costs models based on the local marginal costs. Two examples that use this methodology are the multi area production simulation software (MAPS1) by General Electric and UPLAN network power model (UPLAN-NPM2) by LCG Consulting. Articles [2,11] highlight equilibrium models to conduct studies for the long-term forecast of electricity prices, justifying that the longer the time scope of the study the less detailed modeling capability and the more significant the response of all competitors. Also for long-term studies in market power analysis and market design, Article [11] refers to simulation and equilibrium models as the best alternative. Table 2.3 demonstrates a variety of studies and divides them according to the electricity market models that have been applied within the authors´ investigations in order to study the listed aims and usages. Clearly equilibrium models are mostly used for long-term studies. Table 2.3: Main Electricity Markets Models and their Usage. Source: Adapted from [11]. Optimisation Simulation Equilibrium Major Use Models Models Models Risk management [35,36] Unit commitment [37,38] Short-term hydrothermal [39] coordination Strategic bidding [40,41] Market power analysis [42] [43–48] Market design [49] [43,47,48] Yearly economic planning [50] [51] Long-term hydrothermal [52–56] coordination Capacity expansion planning [57,58] Congestion management [59–64] To model the dynamics of the electricity markets on the long-run Articles [31,65] developed methods based on system dynamics (SD) model. These articles argue that the traditional and few models that assess the long-term market development are not sufficient since they do not take the 1http://www.geenergyconsulting.com/practice-area/software-products/maps 2http://www.energyonline.com/products/uplane.aspx 14 State-of-the-art existence of feedback and system time constants in consideration. Moreover the system evolution is viewed by these models as a sequence of stable and optimal long-run equilibrium states [65] and can lead to excess of investments and over-capacity or the opposite, for instance. “In general terms, the methodology of SD is based on identifying the structure of the system and the logic of the inter-relationships among the different system components to derive its dynamical response. Mathematically, this results in the formulation of the differential equations that represent the system behavior.” [65, chap. 3.Model Description] The study of electricity prices is also related to the study of fuel costs. Articles [66,67] are two examples that have studied this relationship.The previous examines the long-run relation and short-run dynamics between electricity prices and coal, natural gas and crude oil prices in the U.S. For this purpose the study uses historical data and a co-integration method. The latter uses time series methodologies to relate electricity market prices in two major U.S. markets and prices of major electricity generating fuels (Uranium, oil, natural gas and coal). From these articles one can conclude that fossil fuel prices have influence on the electricity market prices and the reverse also happens, but these influences depend on the market studied. The relations can differ depending for example on which type of commodities are more present in the market and their fuel resource and on whether they work at base or peak level. Fossil fuel prices play an important role in electricity markets mainly because they affect production costs of commodities. So forecasting these prices is very useful to project electricity generation costs. Article [68] present a method to forecast oil, natural gas and coal prices in a range of ten years (2008 to 2018). The authors firstly present a review on methods used to estimate commodity prices in the future, which include Geometric Brownian Motion, mean reversion, stochastic price forecasting model and mean reverting jump diffusion model. Then a new model, called long-term trend reverting with jump and dip diffusion, is presented to accomplish the objective of fossil fuel price forecasting. As the objective of this dissertation is not focus on this matter, a detailed explanation of this models is not given. More preeminent is one of the conclusions of the article regarding the prediction results that says that real and nominal oil and natural gas price plus nominal coal price have an increasing trend, while real coal price has a decreasing trend in the future. 2.3 The LCOE Analysis For the model that will be followed within this research it is necessary to use a tool to assist the comparison of the various power generation technologies´ costs. Academic papers and reports on this subject (see [3,14,69–72]) often use a common metric called levelized costs of electricity (LCOE). “The “levelized cost” of supplying electricity using a particular generating technology is a measure of the real total (capital plus operating cost) life-cycle cost per MWh 2.3 The LCOE Analysis 15 supplied.” [73, chap. Comparing Economic Values of Intermittent and Dispatchable Technologies ] Article [70] divides the LCOE calculation in two models: (1) the EGC Spreadsheet model widely used in research reports by the OECD and IEA/NEA and (2) the System Advisor Model (SAM)developed by National Renewable Energy Laboratory (NREL). [71] structures the EGC Spreadsheet model in three parts. The first contains five basic modules(identification, basic assumptions, questionnaire information, generating costs and lifetime generating costs) that provide the reader with the input/output information of the model. Second part includes models for calculating the fuel, CO2and co-generation costs and third part contains two discounting schedules. SAM is a performance and financial model that assists people involved in the renewable energy industry decision-making. The analysis presented in [3] and also the article [72] use an approach based on a discounted cash flow (DCF) analysis for the calculation of LCOE and its future evolution for a certain period. This method is based on discounting financial flows (annual, quarterly or monthly) to a common basis, taking into consideration the time value of money [3]. The LCOE formula is based on Equation 2.1. This simplified equation can evolve into a more complex one depending on the components selected by the responsible of the study and also on the target of it, a specific project or a global one for example. Figure 2.4 presents some possible components that can integrate the LCOE equation. They are divided in three major criteria including investments, annual operations and financial analysis. An example to illustrate each criteria could be costs from project development for investments, fuel costs for operation and inflation for financial analysis. This figure was constructed regarding the reviewed literature on this matter ,mostly cited in this chapter. LCOE =Li f e cycle costs Energy production during li f etime (2.1) Table 2.4 shows some examples taken from the literature that has been reviewed for the LCOE equation. It is possible to see that all presented equations have equal components, such as investments, lifetime and operational and maintenance costs, but each one has is own way to represent them. It is possible to conclude that the LCOE formula is used in the same way by all articles read, namely in its basic equation, yet the aim of its use can be more or less complex. As an example, the equation used in [72], has special components that are used because the article aims to establish an analytical model for solar PV and CSP electricity costs. As seen, the LCOE has several considerations. It is important to notice that these considerations also depend on electricity generation technologies.For example when dealing with renewable technologies it is clear that no fuel cost exists. In Section 2.4 an analysis of actual and past data and perspectives of electricity generation technologies is conducted. Another subject seen in the reviewed literature is the importance of how to estimate the evolution of LCOE in a certain time interval. Both [72] and [74] refer to this issue. The first article 16 State-of-the-art Figure 2.4: Possible Components of LCOE Equation. characterises the reduction in the production costs using learning or experience curves. “...learning curves,or experience curves, which have become a powerful and widely used tool for projecting technological change [75–77], since they describe past evolution of the cost of the systems as a function of the cumulative installed capacity.” [72, chap. Future evolution of the LCOE using the learning curve approach ] The studies [14,71] also call upon this tool to predict the future. The second article refers that, in reality, input parameters regarding costs and also energy production are uncertain and so by using probability distributions for this parameters and Monte Carlo simulation, one can build a LCOE output distribution that allows capturing this uncertain inputs. This approach was also performed in [78]. From the literature review made on this subject it is clear that the LCOE metric is largely applied to compare the different electricity generation technologies. However the use of LCOE for this purpose is said to be insufficient by some articles or companies. Exemplifying, Siemens3come up with a new metric called Society’s Cost of Electricity (SCOE) which takes into account not only the components of LCOE but also other components that can be seen as externalities: Subsidies, transmission costs, variability costs, geopolitical risk impact, environmental impact, social effects and employment effects. This calculation model was created to show that the technology of wind offshore is competitive when compared to other electricity generation technologies. Article [79] refers also to the deficit of the LCOE analysis, by saying that it ignores the variability and integration costs. To surpass the latter, the authors present a new concept called System LCOE which contains generation costs and integration costs. The innovation of this method is in the definition of the integration costs which is directly linked to economic theory. 3http://www.energy.siemens.com/hq/en/renewable-energy/wind-power/SCOE.htm 2.4 Electricity Generation Technologies: Data and Perspectives 17 Table 2.4: Examples of LCOE Equation. Source Equation Comments I0: Investment expenditures in Euro ; At: Annual total costs in Euro in year t ; [14]LCOE =I0+∑n t=1At (1+i)t ∑n t=1 Mt,el (1+i)t Mt,el:Produced quantity of electricity in the respective year in kWh ; i: Real interest rate in % ; n: Economic operational lifetime in years; t: Year of lifetime (1, 2, ...n) . It: Investment expenditures in the year t ; Mt: Operations and maintenance expenditures in the year t; [3]LCOE =∑n t=1It+Mt+Ft (1+r)t ∑n t=1Et (1+r)t Ft: Fuel expenditures in the year t; Et: Electricity generation in the year t; r: Discount rate; n: Life of the system. Et: The amount of electricity produced in year “t”; (1+r)−t: The discount factor for year “t”; It: Investment costs in year “t”; [71]LCOE =(∑tIt+O&Mt+Ft+Ct+Dt)∗(1+r)−t ∑tEt∗(1+r)−t O&Mt: Operations and maintenance costs in year “t”; Ft: Fuel costs in year “t”; Ct: Carbon costs in year “t”; Dt: Decommissioning cost in year “t”. C: Cost of the system; L: Cost of the required land; OPEX: Operation and maintenance costs; I: Insurance costs; [72]LCOE =C+L+∑N n=1(OPEX+I)∗C (1+r)n ∑N n=1S∗TF∗η∗(1−d)n (1+r)n r: Discount rate; S: Available solar resource; TF: Tracking Factor; η: Performance data; d: Annual degradation rate; N: Economic life time of the system; 2.4 Electricity Generation Technologies: Data and Perspectives Studying electricity generation costs and projecting their future evolution implies the knowledge of the state of various technologies, how they are integrated in systems and their state of maturity. In this section data is presented to understand these points. Energy is produced from several technologies which use distinct resources. During the last decades an evolution from the conventional utilities to a system with high penetration of renewable sources has occurred. This situation will most likely continue to evolute in response to the ambitious targets of several world governments to reduce CO2emissions, external independence and 18 State-of-the-art Figure 2.5: Electricity Generation and Population Growth. Source: World Bank(2014), IEA(2014a) and IRENA(2014a). increase efficiency. Figure 2.5 clearly shows this statement, forecasting that in 2030 hydropower and other renewables will have a global share of 34% of the generation mix. Renewable energy sources (RES) include various type of technologies. Figure 2.6 illustrates the evolution of the latter from 2000 to 2013. Here it is possible to see that wind energy is nowadays one of the most advanced and integrated RES. Figure 2.6: Global Renewable Power Generation Cumulative Capacity.Taken from [3]. 2.4 Electricity Generation Technologies: Data and Perspectives 19 Figure 2.7 shows the global growth of wind power capacity. From 2004 this type of technology grew almost 25% per year on average, being United States and China the biggest contributors in the last years. Wind energy can be onshore or offshore. The values of LCOE for these two possibilities differ as is to be seen in Figure 2.8, having offshore higher costs due to the sheer magnitude of the projects [5]. This figure also shows that LCOE values are stabilising, a fact that proves the high maturity of this technology. The costs of wind turbines have decreased significantly since 2007 as proven by Figure 2.9. Figure 2.7: Global Cumulative Growth of Wind Power Capacity.Taken from [4]. Figure 2.8: Levelised Cost of Wind Electricity over Time, developed Market Average (USD/MWh, 2013USD/EUR 0,75). Taken from [5] (Source: Bloomberg New Energy Finance). 20 State-of-the-art Figure 2.9: Cost Trend of Land-based Wind Turbine Prices, by Contract Date.Taken from [4]. Figure 2.10: Global Cumulative Growth of PV Capacity. Taken from [6]. By using the energy of the sun, solar technologies are another possibility of producing renewable energy being solar photovoltaic (PV) and concentrated solar power (CSP) currently the main technologies available throughout the world. PV industry has faced a massive change since 2009, with considerable increases in manufacturing capacities and the move of module manufacturing from Europe and US to China [6]. Since 2009 module prices saw a sharp decrease as seen in Figure 2.11 followed by technology improvements. Decentralised solar PV systems LCOE have become lower than the variable portion of retail electricity price in 2013 in various countries , reaching grid parity (Figure 2.12). This situation provides an incentive to electricity customers to build PV systems and become serious players in electricity retail markets. Since 2009 cumulative CSP capacity has been growing, being Spain and the United States the biggest players of these technologies(Figure 2.13). The main predominant CSP technologies are parabolic troughs, linear Fresnel reflectors (LFR) and towers. The costs of these technologies were expected to decrease, however market opportunities have diminished and the cost of materials increased leading to a slowly diminution of prices [7]. Figure 2.14 shows the LCOE evolution from 2009 to 2013 (Q2). 2.5 System Costs with Renewable Energy 27 Figure 2.19: Number of Countries with Renewable Energy Policies by Type. Taken from [13] receive the same amount of certificates per generated unit of electricity, whereas if they are differentiated some technologies receive more certificates than others. This instrument is also very used around the world as shown in Figure 2.19. 2.5.3 Net Metering Net metering schemes encourage small-scale renewable energy developments, particularly local distributed generation, which allow customers to offset their electricity consumption by injecting renewable energy into the grid [88].However, investors and developers face market risk when facing this scheme, because electricity is usual unknown and the income can only be estimated [86]. As a matter of fact, during the last years net metering usage has increased. (Figure 2.19). 2.5.4 Tenders Tendering schemes were used as a primary policy in the past, yet nowadays they are used in combination with other policy types. The process of tender according to Article [81], consists in calls for tenders by the responsible authorities for specific projects with a certain amount of capacities. Then investors give their bid for the required support level and other specifications (e.g. specific timing of the project, grid positioning and environmental impact) and compete to win the possibility to develop the project. The bid that offers the best conditions wins the tender. Article [81] also divides the tender process in tenders for fixed feed-in tariffs and for target-price feed-in tariffs. 2.5.5 Obligation/mandate and Utility Quota Obligations Another way to promote renewable energy sources in electricity generation is the use of policies that define minimum shares of generations based on these sources. Such schemes can be applied to a group of renewable sources or to a specific source. These policies only define the minimum 28 State-of-the-art share of renewable energy generation, they neither enhance returns nor lower risk, so investors and developers are mainly exposed to market risk [86] 2.5.6 Fiscal Incentives Fiscal incentives can be categorised as shown in Table 2.7. Capital subsidies, grants or rebates provide renewable energy developers with direct cash incentives. Investment grants are granted in the form of non-reimbursable payments at the construction phase of a project and so the generated energy from the project is not directly targeted [81]. These type of policies reduce the upfront costs as well as the LCOE [74]. Energy production payment is a direct payment from governments to one unit of renewable energy generation, while tax credits (applied in production or investment) are an annual income based on the amount of money invested or the energy generation during a certain period [89]. These policies like the above-mentioned ones effectively reduce the LCOE, being tax incentives the most efficient ones [86]. 2.5.7 Public Financing The use of public financing is often made by using loans or investments. Loans consists in financing provided in return of a debt and investments in return of an equity ownership interest [89]. Public competitive bidding is made by governments who provide subsidies to private investors trough a competitive bidding process. The aim of this scheme is to foment competition to obtain the minimum subsidy possible. This process compels investors to develop and improve renewable energies. The subsidy is offered in the form of a feed-in tariff , which results in the establishment of a long term contact. Usually price is the most important factor, therefore market risk is most likely faced by developers and investors [86]. 2.5 System Costs with Renewable Energy 29 Table 2.7: Renewable Energy Support Policies by Country( EU-28, Norway and Switzerland)5Source: Adapted from [13]. 5Spain removed FIT support for new projects in 2012. Incentives for projects that had previously qualified for FIT support continue to be revised. 30 State-of-the-art Chapter 3 Relevant Background Information for the Study The analysis of electricity generation costs in the past and the forecast of its future evolution requires research of several topics. Once this dissertation pretends to study the case of Portugal, a variety of data from historical values of electricity generation capacities and production to electricity imports and exports, as well as on electricity consumption in Portugal needed to be collected. In order to be able to define a research methodology, the costs for different technology plants needed to be searched for and subsequently also presented at this point. Moreover, information regarding the costs derived from the support given to the different electricity generation costs is exposed. Most of this data was retrieved from the web-pages of the national transport grid operator (REN), the energy services regulatory authority (ERSE) and the general directorate for energy and geology (DGEG) - in cases other sources have been used to obtain the exposed information it gets referred to in the following chapter. 3.1 Historical Data Regarding Electricity Consumption and Production in Portugal This section presents necessary collected information on electricity in Portugal. In the past, electricity consumption was very predictable and growing about 3% to 4% per year. However, in the last years, mainly due to economic crisis, the reality is a different one as shown in Figure 3.1.This trend affects national plans for generation capacities and production mix. Besides, electricity generation in Portugal is divided into different types of technologies. Hydropower and fossil fuel plants that use natural gas, coal and oil were the main electricity producers in the past as seen in Figure 3.2. 31 32 Relevant Background Information for the Study Figure 3.1: Evolution of Portugal Electricity Consumption since 2000. Figure 3.2: Electricity Production in Portugal Distributed by its Sources in 2002. Since the year 2000 installed capacity in Portugal has changed drastically with great focus on renewable energy sources (RES). In case of the fossil fuel power plants, all oil power plants have been closed in mainland Portugal and no additional coal plants have been installed. Only in the case of natural gas, investments for new combined cycle centrals have been made. Figure 3.3 shows the evolution of the conventional fossil fuel plants capacity. Figure 3.3: Capacity Evolution of Fossil Fuel Power Plants in Mainland Portugal since 2000. 3.1 Historical Data Regarding Electricity Consumption and Production in Portugal 33 Another way of producing electricity by using fossil fuels is known as "combined heat and power (CHP)" which is also frequently named "co-generation" in Portugal. This type of generation is comprised in a regime called special production regime (PRE). A feed-in tariff, whose definition was reviewed in Section 2.5.1, supports smaller electricity producers so that the projects are economically feasible. Figure 3.4 illustrates the evolution of CHP in Portugal. Figure 3.4: Capacity evolution of Combined Heat and Power in Portugal since 2000. Source: [8]. Hydropower is one of the oldest and most important power plant technologies in Portugal with the first centrals installed more than 50 years ago. This electricity source is often characterised by the power capacity of the central as large hydropower (LHP), when its value is larger than 10 MW, or as small hydropower (SHP), for values under 10MW. It can also be characterised by its infrastructure as run-of-river or dams. SHP capacity stagnated since 2009 around 450 MW ( Figure 3.5).This type of centrals are also included in the PRE regime. Figure 3.5: Capacity Evolution of Small Hydropower (<10MW) in Portugal since 2000. 34 Relevant Background Information for the Study Large hydropower has the greatest capacity installed in Portugal. It is a very important technology once it can provide a great amount of electricity in baseload or peak-demand periods using a renewable and clean primary source. Most of its capacity was installed before 2000, yet since this year until today it has increased 1000 MW: Figure 3.6: Capacity Evolution of Large Hydropower (>10MW) in Portugal since 2000. Another important aspect of the hydropower plants is the possibility to have reversible pump turbines that allow centrals with storage to pump water from the lower reservoir to the higher one. For the electrical system, the energy used to pump water is similar to consumption, therefore one must take this fact into consideration on studying its effects. Figure 3.7 shows the evolution of the energy consumed for pumping. Figure 3.7: Evolution of Portugal Electricity Consumption for pumping since 2000. More recently, renewable energy technologies supported by the PRE regime have been developed in Portugal. One of these renewable energy technologies is the generation of electricity by using wind, which has the greatest growing rate since 2000 regarding its installed capacity, reaching almost 5000MW today as demonstrated in Figure 3.8. 3.1 Historical Data Regarding Electricity Consumption and Production in Portugal 35 Figure 3.8: Capacity Evolution of Wind Onshore Power in Portugal since 2000. Wind power plays a very important role in the electrical energy system, being the greatest provider among the PRE technologies. Figure 3.9 separates the supply by technology to reach the demand on a typical day of autumn. It is a good example of the great contribution of wind power. Figure 3.9: Load Curve Diagram for a Typical Day in Autumn in Portugal.Source: Taken from [9]. Biomass centrals are another electricity generation technology present in Portugal. These types of centrals usually include the ones that are using municipal solid waste (MSW), biogas, crop or forest residues and black liquor. There are two MSW centrals in Portugal that were installed before 2000, one in Lisbon (Valorsul) with 50 MW of capacity and another one in Porto (LIPOR) with 29 MW. The other biomass centrals were installed mainly after 2000. Like CHP, biomass 36 Relevant Background Information for the Study power provides base load capacity. The installed capacity has stagnated in the last years reaching around 700 MW: Figure 3.10: Capacity Evolution of Biomass Power in Portugal since 2000. Furthermore, the conversion of sunlight into electricity, by using solar photovoltaic panels has great potential in Portugal. However, only in 2009 (Figure 2.11) module prices started to decrease to levels that made investments in this field in Portugal viable and that is also why the installed capacity has been lesser than 100 MW until then. During the last years this capacity has increased as seen in Figure 3.11 but it still has a long way to go to reach its full potential. Figure 3.11: Capacity Evolution of Solar Photovoltaic in Portugal since 2000. It is important to notice that the capacity showed in this figure also includes small electricity suppliers. One measure to increase the capacity of solar PV has been made by the Portuguese government who created a law (DL Nº153/2014 [90]) that allows the production of electricity for 3.3 Costs with Electricity Production Incentives in Portugal 43 e/MWh for O&M costs. These amounts were defined by consulting document [97]. 3.3 Costs with Electricity Production Incentives in Portugal As already mentioned, in Portugal exists an incentive scheme based on feed-in tariffs that supports producers included on PRE. However these centrals are not the only ones comprised with this type of mechanisms, but there are also some ordinary centrals that still have supportive contracts, such as CCGT, coal and large hydropower plants. Basically speaking there are three types of contracts in Portugal: CAE, an energy acquisition contract, CMEC, a mechanism to conserve the contractual balance and "power capacity payment", that gives a compensation in the first years of operation to the eligible centrals. Since 2003 CAE contracts have been discontinued, existing only two centrals that remain with CAE: a coal power plant with 576 MW("Tejo Energia") and a CCGT power plant with 990 MW ("Turbogás"). CMEC were then created to replace CAE contracts and "power capacity payments" have been introduced and exist since 2011. These supportive schemes basically result in additional costs that are reverberated on electricity tariffs and subsequently payed by customers. The value of the total costs with this schemes results of adding the additional cost to the market spot price. Figure 3.15 illustrates the evolution of electricity spot market prices in Portugal since 2009 and Table 3.8 depicts the additional costs by the different type of incentives. Figure 3.15: Monthly averages of Electricity Spot Market Price in Portugal since 2009. The amounts of CMEC before 2011 were not found and the value of "power capacity payment" in 2015 is 0 because the incentives will only be incorporated in the 2016 tariffs. This happens due to the governmental decree nº215/2012, of 20 of August, which refers that this amount will only be considered in the year after the financial supportive programme in Portugal. The above-presented information is relevant in order to be compared to the LCOE results that will be demonstrated and explained in Chapter 6. This comparative analysis allows one to understand and to make conclusions of the past and future additional costs of the system. 44 Relevant Background Information for the Study Table 3.8: Annual Electricity Spot Market Prices and Additional Costs of PRE Technologies, CAE, CMEC and "Power Capacity Payment". Chapter 4 Methodology As the principal aim of this dissertation is to study electricity generation costs from the past and project them until 2030, the methodology applied is based on the calculation of the levelized cost of electricity (LCOE), a tool which is commonly used for this type of analyses as reviewed in Section 2.3 and that best fits the purpose of this investigation. This chapter with its respective sections firstly focuses on explaining the LCOE formula that has been applied in this investigation and subsequently describes how the LCOE evolution was calculated for each technology, as well as for the system per se. All the calculations and analysis were made recurring to Microsoft Excel®2010, where a simple and automatic model was designed to facilitate the study of electricity generation costs in the different scenarios that will be explained in Chapter 5. 4.1 LCOE Equation As already mentioned in Section 2.3, the LCOE is the cost of producing one unit of electrical energy. When calculating LCOEs, one needs to consider all the costs and energy produced during the lifetime of the central. The costs are usually divided in capital expenditure (CAPEX), which includes investments, and operational expenditure (OPEX), such as maintenance costs or fuel costs. Considering these facts and having as base the NREL formula1for the LCOE calculation, the equation used in this research is the following: LCOE (e/MWh) = CC ∗CRF +f ixO&M CF ∗8760 +varO&M+f uel +CO2,(4.1) whereCC is the overnight capital cost, CRF is the capital recovery factor, FC is the capacity factor, fixO&Mare the fixed operation and maintenance costs, varO&Mare the variable operation and maintenance costs, f uel are the fuel costs and CO2are the emission costs. 1http://www.nrel.gov/analysis/tech_lcoe_documentation.html 45 46 Methodology The units of CC and fixO&Mare in e/kW and varO&M,f uel and CO2are in e/MWh. The values of these components have already been summarised in Tables A.4 to A.6. One of the components of equation 4.1 is the capital recovery factor which converts present value into a stream of equal annual payments over the lifetime of the central at a specified interest rate. It is given by Equation 4.2. CRF =((1+i)n)∗i ((1+i)n)−1,(4.2) where iis the interest rate and nis the lifetime of the central. According to the reviewed literature, the interest rate used to calculate the CRF can vary depending on the country and technology under consideration. However, in this study an equal interest rate (8%), equivalent to the medium interest rate of the energy sector in Portugal, is assumed for all electricity generation technologies. Another component of the LCOE equation is the capacity factor .The latter is a crucial part of the equation because it defines the number of hours the centrals work per year. In other words it represents the quantity of energy produced by the central in one year. Typical capacity factors of RES centrals in Portugal, excluding large hydropower, were assumed for the LCOE calculation. These values, presented in Table 4.1, were obtained by using historical amounts of energy produced and installed capacity and were calculated by using the following equation: CF(%) = Energyproduced(MWh) Capacity(MW)∗8760 ,(4.3) where 8760 is number of hours in one year. One can notice that in Table 4.1 the capacity factors of wind onshore and solar PV have two values. This happens because in last years the ratio between the energy production from these sources and its installed capacity has raised when compared to the typical values of past years. Table 4.1: Capacity factors of RES(excluding large hydropower). Type of Central Capacity Factor Small Hydropower 26% Wind Onshore 23% (until 2011),28% (after2011) Solar PV 16% (until 2010),18% (after 2010) Biomass and CHP 50% Geothermal 90% In the case of large hydropower plants, the capacity factor values were selected by a different procedure. In the same way specific capital costs of each installed central are known, so is the specific expected production (full list of expected production for each central in Appendix A.1). Once knowing these amounts it is then possible to calculate the capacity factor with Equation 4.3. In case several centrals were installed in the same year, a medium value for the capacity factor 4.1 LCOE Equation 47 was assumed. Figure 4.1 shows the CF values assumed for large hydropower plants in the years that new capacity addition occurs. When observing this figure, the existing difference between the several capacity factors´ centrals become clearer. Figure 4.1: Capacity Factors of Large Hydropower Plants by Installation Year. Another procedure was conducted in order to know the past and future values of CF from coal and natural gas utilities. On the one hand, the historical capacity factors were calculated using Equation 4.3 being its evolution represented in Figure 4.2. The latter figure demonstrates that natural gas utilities have been producing minimum amounts of energy in the last years. Figure 4.2: Historic Capacity Factors of Coal and Natural Gas Utilities. On the other hand, the future CF values of coal and NG utilities were calculated depending on the electric energy they will produce. In order to know the amount of electric energy produced by these centrals, three methods were applied. The reason for this appliance is the need to allow a comparison between more and less electricity production from coal and natural gas utilities. The three methods apply Equation 4.4 to calculate the energy produced by both types of utilities, yet they differ in the way this energy is distributed within each one. 48 Methodology Et=C+Pump −ERES +(Imp −Exp),(4.4) where, •Etis the total energy produced by coal and NG utilities; •Cis the energy consumption; •Pump is the energy consumption used by hydropower plants to pump water •ERES is the total energy produced by RES power plants •Imp is the energy imported from Spain; •Exp is the energy exported to Spain. All components of this equation are in GWh. It is important to refer that the total energy produced by RES utilities present in Equation 4.4 was obtained by adding the energy produced by each RES power plant, which was calculated by Equation 4.5, that use the typical capacity factors presented in Table 4.1. Moreover, the capacity factor of large hydropower plants used to calculate its total energy production was the same as for SHP plants. Energy produced (GW h) = CF(%)∗Capacity(MW)∗8760 1000 (4.5) The above-mentioned methods for the calculation of the energy produced by coal and NG power plants are: 1. The energy produced by the coal utilities follows the last year trends, where the capacity factor values stayed around 75%. For NG power plants, the energy produced is given by the following equation: ENG(GWh) = Et(GW h)−Ecoal(GWh),(4.6) where, •ENG is the total energy produced by NG power plants; •Etis the total energy produced by NG and coal utilities, obtained by using Equation 4.4; •Ecoal is the total energy produced by coal power plants. 4.2 Method of Calculation of the LCOE Evolution 49 2. The energy produced by the coal utilities is proportional to its installed capacity and given by Equation 4.7. The energy produced by NG power plants is given by Equation 4.6. Ecoal(GWh) = Et(GWh)∗Pcoal(MW) Pcoal(MW)+PNG(MW),(4.7) where, •Pcoal is the power capacity of coal utilities; •PNG the power capacity of NG utilities. 3. In this case the amount of energy assign to NG power plants is the majority of the total energy that these two thermal power plants can produce. However one must take into consideration that for security reasons of the electrical system, regarding the operating reserve, coal utilities must stay with capacity factors above 5%. 4.2 Method of Calculation of the LCOE Evolution Having exposed the LCOE calculation used within this research in the last section, it is possible at this point to explain the manner the LCOE evolution will be calculated for each technology, as well as in case of the system LCOE evolution from 2000 until 2030. To calculate the LCOE for each type of power plant and its evolution, it is firstly required to calculate the LCOE for each year in which new capacity was and will be installed. This LCOE is then extended through the lifetime of the central or at a maximum until 2030. Afterwards an overall LCOE (OLCOE) is calculated for each year, which is a weighted average by its installed capacity (Equation 4.8). OLCOEx,i(e/MWh) = n ∑ i=2000LCOEx,i∗Pinst,i n ∑ i=2000Pinst,i ,(4.8) where, •OLCOEx,iis the overall LCOE of the power plant technology x in year i; •LCOEx,iis the LCOE of the power plant technology x in year i; •Pinst,iis the installed capacity of power plant technology x in year i; •n= i + lifetime of power plant technology x. Finally, it is possible to calculate the system LCOE (SLCOE) evolution by using the OLCOE results for each power plant technology. This calculation( Equation 4.9) follows the same principle 50 Methodology as the calculation of the OLCOE, however in this case the weighted average is calculated by the energy produced by each power plant technology. SLCOEi(e/MWh) = ∑ xOLCOEx,i∗Ex,i ∑ x Ex,i ,(4.9) where, •SLCOEiis the system LCOE in year i; •OLCOEx,iis the overall LCOE of the power plant technology x in year i; •Ex,iis the energy produced by the power plant technology x in year i. Chapter 5 Scenarios In order to enable the projection of electricity generation costs and the comparison between the different types of electricity generation technologies, a scenario needs to be created to simulate the evolution of installed capacity, electricity consumption, and electricity imports, as well as exports. Moreover, to compare the costs of different solutions in which installed capacity of some electricity generation technologies are changed, other scenarios need to be developed. Therefore, this chapter explains the basic scenario that has been created in the first place, and subsequently clarifies how further scenarios have been developed and assumed on the basis of the first one. 5.1 Base Scenario The creation of this scenario was based on the presuppositions of the Portuguese national action plan for renewable energies (PNAER) and national plan for climate changes (PNAC 2020/2030) [98,99]. The PNAER was executed by the Portuguese government as a result of the European directive regarding the promotion of the use of energy from renewable energy sources [82], in which the European commission established that member states are to establish national action plans which set the share of energy from renewable sources. Both PNAER and PNAC, that was available for consultation more recently, make projections for installed capacity and consumption, such as: • The structure of the electricity consumption does not suffer significant changes, maintaining a structure similar to 2010; • The installed capacity, on a reference case, grows 4,1 GW in 2020 when compared to 2010 values. This amount results from the shutdown of the coal central in Sines, the addition of two new CCGT centrals and the addition of wind onshore, hydropower and solar PV capacity, allowing a share of 63% of renewables in the production of electricity; 51 52 Scenarios • For 2030 the installed capacity will raise 5.4 GW when compared to the capacity of 2010, which implies that the capacity levels of wind onshore will be 6.1 GW, solar PV 1.2 GW and hydropower 9.1 GW having coal utilities disappeared; • In 2030, renewable sources will assure 62% to 70% of the electricity produced depending on the consideration of more or less energy produced by natural gas utilities. Moreover this document also includes, in the reference case, energy efficiency objectives which can lead to a lower electricity production in 2030 from 7% to 18 % when compared to 2010 levels. As already mentioned before, the creation of the base scenario takes into consideration the above-described presuppositions, but will also take notice of the actual reality of the electrical system and the perspectives obtained by professionals in the area. The following sub-sections describe the projections assumed in the base scenario. 5.1.1 Capacity Projections It is firstly important to refer that the projections made in this chapter refer to future years starting in 2015, however in most figures values are shown since 2010 as a means of comparison. Starting with future projections of large hydropower plants, in 2007 a national plan regarding the high potential of hydroelectricity was elaborated [100]. This plan defined the investments that would take place until 2020, in order to achieve the adequate levels of the hydroelectric potential. The past years showed that most of the initially defined capacity additions have not taken place and are not going to take place in the designated years of the plan. Therefore, the projection made for new installed capacity of large hydropower plants follows a different trend in this research and is illustrated in Figure 5.1 (full list of new large hydropower plants can be found in Appendix A.1). This figure shows that there is a strong addition of large hydropower capacity, reaching 8.77 GW in 2030. Figure 5.1: Projected Capacity of Large Hydropower (>10MW) in Portugal until 2030 (Base Scenario). 5.1 Base Scenario 59 during the year( this also occurs between Spain and France), due to unavailability of transmission lines and other elements of the network. The annual reports of REN ( [103,104]) provide the annual export capacities, as well as the frequency of its occurrence. In 2012 and 2013, the export capacity interval with major occurrence was located in between 2000 and 2200 MW. When observing figure 5.12, it is possible to verify that the distribution of the export and import quantities suffers major distortions on these levels. Additionally, in 2013 there has been a strong capacity limitation occurrence for 1300 MW, which explains the accentuated distortion that can be verified within these amounts in the respective histogram. In order for this effect not to occur, an increase of the capacity of interconnection is necessary. In case this limitation, which has been verified during the last years until today, would be overcome it would then also be possible to expect that the export values would increase 10%. Concerning the capacity of imports, the histogram of Figure 5.12 shows that there are no limitations, since the distribution is not distorted. It is then not expected that import values will suffer great changes in the future. Figure 5.12: Histogram of Electricity Imports/Exports in Portugal.Data from 2012 and 2013. By considering the above-mentioned facts, as well as the interconnection capacity between Portugal and Spain which is expected to augment to 2800MW in 2016 and to 3000MW in 2017, while that between Spain and France is also expected to increase to 2800MW in 2016 [105], the projection illustrated in Figure 5.13 has been realised based on/ concerning the import balance in Portugal. 60 Scenarios Figure 5.13: Projected Balance of Imports in Portugal until 2030. 5.2 CHP Scenario For the conduction of this dissertation it has been highly relevant to study the effects of diverse power generation technology solutions on the system costs. That is why other scenarios that introduce small modifications to the base scenario needed to be developed. In fact, one of the scenarios created projects a different trend for the evolution of CHP power plants. Contrarily to the assumption that the capacity of CHP centrals will decrease in the future, this scenario considers that the following two further situations may occur: • The first one, named "CHP remains", defines that no future investments will happen but the actual capacity will remain operational, which represents an addition to the total capacity of PRE power plants when compared to the Base Scenario (see Figure 5.14); • The second case considers the law allowing electricity production for auto-consumption [90], which means that the energy produced by these centrals is included in the consumption, decreasing its value when compared to the base scenario (see Figure 5.15). In fact this measure can be considered as an efficiency measure, since it decreases the consumption. That is why this scenario is called "CHP as efficiency". 5.3 Solar PV Scenario 61 Figure 5.14: Total Projected Capacity in Portugal until 2030 ("CHP remains" Scenario). Figure 5.15: Projected Consumption Including Energy for Pumping in Portugal until 2030 ("CHP as Efficiency" Scenario ). 5.3 Solar PV Scenario In case of the Solar PV Scenario, the modifications made to the base scenario are the following: • Taking into consideration the same fact explained in the "CHP as efficiency" scenario, regarding the law of own consumption, and that the module prices of solar panels have reached levels that made investments in this field viable [20], it is projected in this scenario that until 2020, 300 MW of Solar PV will be installed for own consumption and that this value will augment to 1200MW by 2030. 62 Scenarios As in the "CHP as efficiency" scenario, the projection in this scenario affects the consumption of electricity, by decreasing its amounts ( see Figure 5.16) and thus is named "PV as efficiency". • This scenario, named "PV More", assumes that the capacity addition of the previous case is not installed for own consumption but has centralised solar PV centrals, which means that the capacity of solar PV power plants increase when compared to the base scenario, reaching 2.4 GW in 2030( see Figure 5.17). Figure 5.18 shows that this presupposition also increases the total capacity of the PRE power plants. Figure 5.16: Projected Consumption Including Energy for Pumping in Portugal until 2030 ("PV as Efficiency" Scenario ). Figure 5.17: Projected Capacity of Solar PV Power Plants in Portugal until 2030 ("PV More" Scenario). 5.4 Wind Onshore Scenario 63 Figure 5.18: Total Projected Capacity in Portugal until 2030 ("PV More" Scenario). 5.4 Wind Onshore Scenario Regarding the fact that onshore wind plants have high levels of penetration and maturity around the world, as mentioned in Section 2.4, the last scenario created changes to the projected installed capacity of onshore wind in the base scenario. In this case the additions to the installed capacity of onshore wind result in almost "6000 MW" of total capacity in 2020 and "8300 MW" in 2030, as demonstrated in Figure 5.19. As a consequence of this addition, the total installed capacity of PRE power plants will also increase, reaching more than "10000 MW" in 2030 (see Figure 5.20). Figure 5.19: Projected Capacity of Wind Onshore Power Plants in Portugal until 2030 ("Wind More" Scenario). 64 Scenarios Figure 5.20: Total Projected Capacity in Portugal until 2030 ("Wind More" Scenario). Chapter 6 Results Once having retrieved the relevant information needed for the study (see Chapter 3) and created scenarios for future developments in the electric power system (see Chapter 5), it is now possible to calculate past electricity generation costs, as well as their future projections by applying the methodology designed and explained in Chapter 4. Furthermore, with the above-mentioned results in mind, it is possible to study the costs that the electric power system had and will have with the production of electricity. Consequently this chapter firstly presents the energy production results by each type of electricity generation technology according to the scenarios that have been previously elaborated (see Chapter 5) and are required for the appliance of the LCOE equation. Subsequently, the results of the past and future LCOEs, the overall LCOEs and the system LCOE are demonstrated in order to end this chapter in a way that it is possible to draw conclusions regarding the effects on the costs that the system has with the supportive schemes for electricity producers . 6.1 Energy production This section summarises the results of the energy produced by each type of technology when regarding the different scenarios mentioned earlier. For an easier reading, the information presented within this section is divided in various subsections that refer to the different electricity generation technologies, while applying the scenarios that have been exposed in the previous chapter and containing the energy produced during these scenarios. It should also be added at this point that in Appendix A.2 are presented the production mixes resultant from each scenario. 6.1.1 Energy Production of Large Hydropower Since large hydropower plants have the same installed capacity in all developed scenarios, the energy produced by this source remains equal. Figure 6.1 shows these values. It is possible to to verify an increase of the energy produced which corresponds to the additions of the capacity installed assumed in the base scenario. 65 66 Results Figure 6.1: Energy Produced by Large Hydropower (>10MW) in all Scenarios. 6.1.2 Energy Production of SHP and Biomass Both biomass and SHP, as in the case of large hydropower plants, have the same capacity in all the scenarios, which is defined in the base scenario. Therefore, the energy produced by this source is equal for all the scenarios (see Figure 6.2) Figure 6.2: Energy Produced by Small Hydropower(<10MW) and Biomass in all Scenarios. 6.1.3 Energy Production of CHP For CHP centrals there are two different situations that have different values of produced energy: The first one is related to the capacity defined in the "CHP remains" scenario and the second one is related to the capacity defined in the base scenario. Moreover, the latter is also used in the remaining scenarios. Figure 6.3 shows the energy produced by CHP centrals in both situations. 6.1 Energy production 67 Figure 6.3: Energy Produced by CHP Centrals in the "CHP remains" Scenario and in the Remaining Scenarios. 6.1.4 Energy Production of Solar PV The energy produced by Solar PV power plants also differs in two ways, depending if it is considered the base scenario or the "PV More" scenario. Since in the "PV More" scenario the installed capacity over the years is greater than in the base scenario, the energy produced in the first has a more pronounced growing rate than the second (see Figure 6.4). Figure 6.4: Energy Produced by Solar PV Centrals in the "PV More" Scenario and in the Remaining Scenarios. 6.1.5 Energy Production of Onshore Wind Another electricity generation technology with two distinct amounts of energy produced, is the Onshore wind, with the respective values represented in Figure 6.5. Again, the growing rate of energy produced is more pronounced in the "Wind More" scenario than in the base scenario. 68 Results Figure 6.5: Energy Produced by Onshore Wind Centrals in the "Wind More" Scenario and in the Remaining Scenarios. 6.1.6 Energy Production of Coal and NG Power Plants In the case of coal and NG utilities there are several situations with different amounts of energy produced, due to the fact that this electricity resources depend on the energy produced by the remaining power plants, as defined in Section 4.1. As a consequence of the above-mentioned fact, the energy produced by coal and NG utilities change in each scenario considered. Moreover, these amounts also depend on the installed capacity that has been considered, defined by cases A,B and C in Sub-section 5.1.1 and on the methods used for their calculation. Figure 6.6: Energy Produced by Coal utilities in the Base Scenario, according to the case and method considered. 6.2 Electricity Generation Costs 75 Figure 6.14: LCOE Projections of Coal Utilities in all the Created Scenarios. Figure 6.15: LCOE Projections of NG Utilities in all the Created Scenarios. Once disposing of the LCOE projections until 2030, as well as of the past OLCOE values it is possible to project the OLCOE of a variety of technologies. Similarly as in the case of LCOE, these values are equal for all technologies in any scenario, except for coal and NG which depend on the energy produced in each scenario. The results of these projections are located in figures 6.16 to 6.18. It is possible to observe in the first figure that only the OLCOE of solar PV technology decreases until achieving 110 e/MWh in 2030. In case of the other technologies, however, the OLCOE remains stable during the years, being wind energy once again the one that presents the lowest values. 76 Results Figure 6.16: Overall LCOE Projections of RES and CHP. Figure 6.17: Overall LCOE Projections of Coal Utilities in all the Created Scenarios. When analysing Figures 6.17 and 6.18 it is once again possible to recognise that coal centrals have the lowest values in case A, whereas NG centrals have the lowest values in case B for the same reasons that have been mentioned earlier. In addition, the existence of a certain kind of disparity regarding higher OLCOE values of various scenarios may be observed which will continue to have effects on SLCOE and will be analysed in the next sub-section. 6.2 Electricity Generation Costs 77 Figure 6.18: Overall LCOE Projections of NG Utilities in all the Created Scenarios. 6.2.3 SLCOE Projections Having explained and demonstrated the OLCOE projections in the previous section, as well as the energy produced by each electricity production technology in the diverse scenarios, it is now possible to present the SLCOE results for each scenario. Contrarily to what happened to the LCOE and OLCOE values in RES and CHP centrals, these values depend on the energy produced and are therefore different in each scenario that has been created. First of all, Figure 6.19 presents the SLCOE projection within the current electric system panorama [106]. For electricity production, carbon prices are lower than natural gas prices and NG centrals work with very low capacity factors. In this scenario the SLCOE should/will already decrease this year, being the peak in 2014 owed to the strong investment in the hydroelectric central “Baixo Sabor” (see Appendix A.1), and stabilise during the next years around 80e/MWh. Figure 6.19: System LCOE Projections in the Actual Panorama. Hereupon the results of SLCOE in all developed scenarios exposed in Chapter 5are presented 78 Results (see Figures 6.20 to 6.23). In these figures it is possible to perceive that in any scenario the SLCOE projections assume lower values in case B and C. It is true that in these cases the NG price trend is inverse that of the base scenario, yet even if this tendency will retain, the results would be similar, because the LCOE of coal centrals is higher, as is to be seen in Figure 6.13. Figure 6.20: System LCOE Projections in the Base Scenario. Figure 6.21: System LCOE Projections in the "CHP remains" and in the "CHP as efficiency" Scenarios. In addition, it gets highlighted that in both CHP and Solar PV scenarios (see figures 6.21 and 6.22) the situations that may differ between installed capacity as centralised power plants or as own consumption, present almost equal results. 6.2 Electricity Generation Costs 79 Figure 6.22: System LCOE Projections in the "PV More" and in the "PV as efficiency" Scenarios. Figure 6.23: System LCOE Projections in the "Wind More" Scenario. When comparing Figure 6.23 to 6.20 it is possible to deduce that the SLCOE values in the “Wind More” scenario are not worse than in the base scenario, which has not only to do with the fact that wind energy has a high installed capacity in Portugal and therefore a strong weight in its energy mix, but also with the fact that the LCOE of this technology is the lowest one within all technologies considered in this research. In order to conclude this section, a comparison between the results with the lowest SLCOE values in each scenario gets presented ( see Figure 6.24. Actually the values are quite similar, yet one may notice that CHP and Solar PV present worse results than the base scenario. In the first 80 Results case this is due to the fact that both the OLCOE as well as the LCOE of CHP are high (see Figures 6.13 and 6.16), while in the second case solar PV penetration is still low and its OLCOE on a level that will not achieve the foreseen LCOE for that technology until 2030. Even more, when comparing these values with those presented in Figure 6.19 it may be noticed that they present lower values, being the difference in 2030 of 10e/MWh. Figure 6.24: System LCOE Projections in the Best Case of each Scenario. 6.3 Comparison of the Results with the Additional Costs of the System As mentioned in the beginning of this chapter, the final aim of this research is to use the OLCOEs and system LCOE results in order to be able to understand and compare them with the additional costs paid by the system to the electricity producing centrals. In chapter 3the historical values of the additional costs from 2009 to 2014 have been exposed, including PRE, CAE, CMEC and “power capacity payment” supports. Additionally the market prices have been presented that originate the total cost of the system in each type of central when being added to the additional costs. For this research it is necessary to dispose of a projection of the above-mentioned prices until 2030. It is foreseen that due to the increase of the interconnection of the European markets, as well as the creation of a single energy market the electricity costs would decrease to the European average which is situated around 40e/MWh. 6.3 Comparison of the Results with the Additional Costs of the System 81 Given these facts, the following projection for the electricity spot market prices in Portugal has been made: Figure 6.25: Projection of the Portuguese Electricity Spot Market Prices until 2030. Figures 6.26 to 6.34 have been compiled by having in mind the above-mentioned, as well as the OLCOE and SLCOE values explained in the previous section. Figure 6.26: Overall LCOE and Total Additional Cost of Solar PV. Figure 6.27: Overall LCOE and Total Additional Cost of Wind Onshore. 82 Results Figure 6.28: Overall LCOE and Total Additional Cost of SHP. Figure 6.29: Overall LCOE and Total Additional Cost of CHP. Figure 6.30: Overall LCOE and Total Additional Cost of Biomass. 6.3 Comparison of the Results with the Additional Costs of the System 83 Figure 6.31: Overall LCOE of Large Hydropower Plants and Total Additional Cost with CMEC. Figure 6.32: Overall LCOE of Coal Utilities and Total Additional Cost with CAE. Figure 6.33: Overall LCOE of NG Utilities and Total Additional Cost with CAE. In these figures it is possible to verify that until 2015 the values of the total cost of each PRE technology are higher than the calculated OLCOE. In fact, this was expected otherwise the investment in these technologies would not be viable. However, in case of the remaining technologies this does not happen. As already explained in Chapter 3, there are certain centrals (excluding PREs) that dispose of CAE or CMEC contracts such as coal or NG centrals in the case of CAE 84 Results contracts and NG or Large Hydropower centrals with CMEC contracts. The comparison between these additional costs and the OLCOE values of the respective centrals permits a better understanding of the fact that they get remunerated below their total costs. Besides, once the additional cost “power capacity payment” has not entered during the past years it has also not been included in this research. Regarding each PRE technology value, one notices that in some cases the additional cost which has been paid to these centrals has not accompanied the decrease of OLCOE, above all in the case of Solar PV (see Figure 6.26). The increase of the additional costs verified in the last years can be explained by the existence of adjustments, which derived from accumulated costs of the previous years. Yet, it was expected that the additional cost followed the OLCOE. In order to be able to make the projections of the additional costs until 2030 it was assumed that they would follow a decreasing trend in PRE technologies, yet never falling to values that would not permit a margin between the OLCOE and the total cost below 20%. As for the remaining electricity generation technologies, to reach this margin and maintain it in future years the additional costs need to increase. The projections for the additional costs until 2030 are represented in the previous figures. These projections result in a total additional cost for the system, which can be calculated with the previously mentioned additional costs and the energy produced by each electricity generation technology in the base scenario (Case A). The projections of the total additional cost of the electric system are presented in Figure 6.34. Figure 6.34: System LCOE (Base Scenario, Case A) and Total Additional Cost of the Electric System. A.2 Energy Produced and Energy Mix 91 Figure A.4: Electricity Production Mix in Portugal in the "CHP as efficiency" Scenario (Case A). Figure A.5: Electricity Production Mix in Portugal in the "PV More" Scenario (Case A). 92 Appendix Figure A.6: Electricity Production Mix in Portugal in the "PV as efficiency" Scenario (Case A). Figure A.7: Electricity Production Mix in Portugal in the "Wind More" Scenario (Case A). A.3 LCOE Components 93 A.3 LCOE Components Figure A.8: LCOE of Coal Utilities Separated by each Component. Figure A.9: LCOE of Natural Gas Utilities Separated by each Component. Figure A.10: LCOE of CHP Power Plants Separated by each Component. 94 Appendix Figure A.11: LCOE of Large Hydropower Plants Utilities Separated by each Component. Figure A.12: LCOE of SHP Plants Separated by each Component. Figure A.13: LCOE of Geothermal Power Plants Separated by each Component. A.3 LCOE Components 95 Figure A.14: LCOE of Wind Onshore Power Plants Separated by each Component. Figure A.15: LCOE of Solar PV Power Plants Separated by each Component. Figure A.16: LCOE of Biomass Power Plants Separated by each Component. 96 Appendix A.4 Capital and O&M Costs Table A.4: Capital Costs in e/kW of the Different Electricity Generation Technologies Relevant for this Study. Source: [19–21]. A.4 Capital and O&M Costs 97 Table A.5: Fixed O&M Costs in e/kW of the Different Electricity Generation Technologies Relevant for this Study. Source: [19]. 98 Appendix Table A.6: Variable O&M Costs in e/kW of the Different Electricity Generation Technologies Relevant for this Study.Source: [19] . References [1] World Energy Council. 2015 World Energy Issues Monitor. Technical report, 62–64 Cornhill London EC3V 3NH, 2015. URL: www.worldenergy.org. [2] J. Lawarree, M. Gallanti, and a. Venturini. State-of-the-art of electricity price forecasting. International Symposium CIGRE/IEEE PES, 2005., pages 110–119, 2005. doi: 10.1109/CIGRE.2005.1532733. [3] IRENA. Renewable Power Generation Costs in 2014. Technical report, International Renewable Energy Agency, 2015. URL: http://www.irena.org/ DocumentDownloads/Publications/IRENA_RE_Power_Costs_2014_ report.pdf. [4] Energy Agency(IEA). Technology roadmap : Wind energy. Technical report, International Energy Agency, 2013. URL: https://www.iea.org/publications/ freepublications/publication/Wind_2013_Roadmap.pdf. [5] World Energy Council. World Energy Perspective - Cost of Energy Technologies. Technical report, World Energy Council, London, 2013. URL: www.worldenergy.org. [6] Energy Agency(IEA). Technology Roadmap: Solar Photovoltaic Energy. Technical report, International Energy Agency, 2014. doi:10.1007/SpringerReference\_7300. [7] Energy Agency(IEA). Technology Roadmap: Solar Thermal Electricity. Technical report, International Energy Agency, 2014. doi:10.1007/SpringerReference\_7300. [8] Cogen Portugal. Accessed: 2015-04-27. URL: http://www.cogenportugal.com/. [9] REN - Rede Eléctrica Nacional. Caracterização da rede nacional de transporte para efeitos de acesso à rede em 31 de Dezembro de 2014. Technical report. URL: http://www.centrodeinformacao.ren.pt/PT/publicacoes/ CaracterizacaoRNT/Caracteriza%C3%A7%C3%A3odaRNT31-12-2014.pdf. [10] Van Beck N.M.J.P. Classification of Energy Models. FEW Research Memorandum;Vol.777, 1999. [11] Mariano Ventosa, Álvaro Baíllo, Andrés Ramos, and Michel Rivier. Electricity market modeling trends. Energy Policy, 33:897–913, 2005. doi:10.1016/j.enpol.2003. 10.013. [12] Edward S. Rubin, Chao Chen, and Anand B. Rao. Cost and performance of fossil fuel power plants with CO2 capture and storage. Energy Policy, 35(9):4444–4454, 2007. doi: 10.1016/j.enpol.2007.03.009. 99 100 REFERENCES [13] REN21. Renewables 2014 Global Status Report. Technical report, Paris:REN21 Secreteriat, 2014. URL: http://www.ren21.net/Portals/0/documents/Resources/ REN21_AnnualReport_2014_web.pdf. [14] Christoph Kost, Johannes N. Mayer, Jessica Thomsen, Niklas Hartmann, Charlotte Senkpiel, Simon Philipps, Sebastian Nold, Simon Lude, Noha Saad, and Thomas Schlegl. Levelized Cost of Electricity Renewable Energy Technologies. Technical Report November, Fraunhofer Institute for Solar Energy Systems ISE, Freiburg - Germany, 2013. [15] Energy Agency(IEA). Technology Roadmap : Hydropower. Technical report, International Energy Agency, 2012. doi:10.1007/SpringerReference\_7300. [16] British Petroleum. BP Statistical Review of World Energy 2015. Technical Report 64th edition, 2015. URL: bp.com/statisticalreview. [17] DGEG and Governo de Portugal. Relatório de Monitorização da Segurança de Abastecimento do Sistema Elétrico Nacional 2013-2030. Technical report, 2013. [18] Exchange rates - OECD Data. Accessed: 2015-04-22. URL: https://data.oecd. org/conversion/exchange-rates.htm. [19] OpenEI-Transparent Cost Database. Accessed : 2015-04-24. URL: http://en. openei.org/apps/TCDB/. [20] Fraunhofer ISE. Current and Future Cost of Photovoltaics. Long-term Scenarios for Market Development, System Prices and LCOE of Utility-Scale PV Systems. Technical report, Study on behalf of Agora Energiewende, 2015. [21] Rui M G Castro. INTRODUÇÃO À ENERGIA MINI-HÍDRICA. Instituo Superior Técnico, Universidade Técnica de Lisboa, 2002. [22] R. B. Hiremath, S. Shikha, and N. H. Ravindranath. Decentralized energy planning; modeling and application-a review. Renewable and Sustainable Energy Reviews, 11(5):729–752, 2007. doi:10.1016/j.rser.2005.07.005. [23] Toshihiko Nakata. Energy-economic models and the environment. Progress in Energy and Combustion Science, 30(4):417–475, 2004. doi:10.1016/j.pecs.2004.03.001. [24] Jean-Charles Hourcade, R Richels, J Robinson, W Chandler, O Davidson, D Finon, MJ Grubb, K Halsneas, K Hogan, M Jaccard, et al. Estimating the costs of mitigating greenhouse gases. Climate Change 1995, Economic and Social Dimensions of Climate Change, Contribution of Working Group II, pages 263–296, 1996. [25] Michael Grubb, Jae Edmonds, Patrick ten Brink, and Michael Morrison. The costs of limiting fossil-fuel co2 emissions: A survey and analysis. Annual Review of Energy and the Environment, 18(1):397–478, 1993. doi:10.1146/annurev.eg.18.110193.002145. [26] Gustavo Brandão Haydt de Souza. A multi-objective decision support methodology for developing national energy efficiency plans. Ph. d. thesis in sustainable energy systems, Faculdade de Engenharia da Universidade do Porto, 2011. [27] Cutler J Cleveland and Christopher G Morris. Dictionary of energy. Elsevier, 2005.